Introduction: Kenya – A democracy in retreat?

Abstract

The post-election crisis of January 2008 brought Kenya close to collapse and the status of a failed state. Following the abrupt proclamation of Mwai Kibaki, the incumbent president, as victor in a highly contentious presidential election, peace was disrupted by severe ethnic violence between supporters of the opposition Orange Democratic Movement (ODM) and the Party of National Unity (PNU). This saw up to 2,000 people killed and as many as 300,000 displaced from their homes. This analysis locates the origin of the crisis in, variously, a background of population growth and extensive poverty; and ethnic disputes relating to land going back to colonial times (notably between Kalenjin and Kikuyu in the Rift Valley). More immediately, what stoked the conflict is the construction of political coalitions around Kenya's 42 ethnic groups, although the 2007 election campaign was critically shaped by ODM's rhetoric of ‘41 against one’ (the Kikuyu); and not least, this survey records the diffusion of violence as a result of elite manipulation of armed militias which, since 1992, have steadily eroded the state's monopoly of violence. While summarising how external mediation and the elite interest in political stability prevented the country falling apart, and led to the formation of a power-sharing government, the analysis proposes that a reluctance by the Grand Coalition partners to undertake fundamental reform of the constitution means that Kenya remains a ‘democracy at risk’, and faces a real possibility of slipping into state failure.

Keywords:

• ethnic violence in Kenya
• post-election violence
• diffusion of violence
• armed militias
• power sharing
• constitutional reform

The horror and mayhem that followed the massively flawed 27 December 2007 general election in Kenya highlighted the violent rollback of many of the democratic gains Africa had made since the early 1990s. The ‘Kenya crisis’ has come to refer to the political and humanitarian catastrophe that engulfed the country after the incumbent president, Mwai Kibaki, was controversially declared by the Electoral Commission of Kenya (ECK) as the winner of the country's closely contested election and sworn in for a second five-year term on 30 December. Violence erupted in strongholds of the Orange Democratic Movement (ODM) headed by Raila Odinga, mainly in Nairobi, Nyanza and the Rift Valley as Odinga and his supporters protested that they had been robbed of victory. The ensuing electoral dispute quickly metamorphosed into a deadly orgy of ‘ethnic’ slaughter, rape and plunder reminiscent of the 1994 Rwandan genocide. On New Year's day, 39 people, mainly women, children and disabled were burnt to death by marauding ethnic supporters of the ODM who set ablaze a church in Eldoret where they had taken asylum. The violence targeted mainly the Kikuyu, Kibaki's ethnic group, and others suspected of voting for his Party of National Unity (PNU). Previous multiparty elections in 1992 and 1997 and to a lesser extent 2002 were also marred by cycles of ethnic violence; but 2007 violence was diffused, further eroding the state's already fragile monopoly over violence. While an overwhelmed Kenyan police managed to maintain peace in many parts of the country, its officers were accused of rape, shooting innocent demonstrators, and other grievous violations of human rights. From late January 2008, Kikuyu youths and associated Mungiki militias, reacting to the killings and displacement of their kith and kin in ODM strongholds, engaged in reprisal attacks in Nairobi and parts of the Rift Valley and Central Kenya against the Luo, Luhya and Kalenjin groups, the ODM's ethnic support base. By the time the violence subsided in April 2008, it had left 1,000 to 2,000 people dead and 600,000 displaced from their homes and communities and cost the economy over 100 billion Kenya shillings (around $1.5billion) (Africa Research Institute 2008).

Although Kenya, East Africa's economic powerhouse, had not become a ‘failed state’, the crisis had pushed it to the brink of collapse and failure. Customarily hailed as a relatively peaceful and stable haven in the turbulent continent, in the post-violence period the country has come to signify the tragic slide from the optimism that greeted the global ‘third wave’ of democracy as it swept Africa in the early 1990s to pessimism and despair (Huntington 1991; Young 1999). Kenya has now firmly entered the ever-growing category of democracies at risk of failure across Africa, the limits and vulnerability of ‘electoral democracy’ having been exposed in the face of stalled or failed constitutional experiments and weak institutions. The country highlights the ever-present risk of election disputes degenerating into deadly conflicts in the context of weak institutions, elite fragmentation, surging ethnic nationalism, authoritarian undertows, corruption, widening social economic inequalities, historically embedded injustices, grinding poverty, debt overhang and the spillover effects of climate change and global recession.

An April 2008 power-sharing agreement between President Kibaki and Raila Odinga, brokered initially by an African Union-sponsored mediation team, headed by Ghanaian President John Kufuor but subsequently backed up a team of ‘eminent African personalities’ led by former UN Secretary-General Kofi Annan, ended the carnage. The peace deal encompassed power-sharing between ethnic-based elite factions and allocation of executive positions to the major groups in a ‘grand coalition government’ involving the parties in dispute, albeit at the cost of the appointment of a 40-member cabinet, the largest (and most expensive) in Kenya's history. However, it is not clear that the ‘elite truce’ will hold and address the fundamental causes of the crisis and prevent Kenya's return to instability and chaos. As Michael Chege (2008, 126) rightly argues, Kenya is still at risk. This is because the April agreement left out important pillars of the consociational theory on which the Kenyan grand coalition government is based, namely autonomy and devolution of power to warring groups, mutual veto, and the equality of vote often guaranteed by the electoral system of proportional representation (PR). As such, analysts have urged a fundamental renegotiation of the foundations of a democratic state in Kenya in the context of the post-conflict reforms as crucial to preventing the recurrence of violence and the risk of state failure. Yet, it is worth noting, as Susanne Mueller (2008, 186) has cogently pointed out, that Kenya's slip to the brink started long before the 2007 election provided the spark for the conflagration. While locating Kenya's 2008 crisis within the immediate circumstances relating to the 2007 election, this contribution attempts to analyse the mayhem within the broader canvas of the contradictions embedded in the country's political economy.

The making of a crisis

Despite capitulating to domestic and international pressure to return Kenya from one-partyism to pluralist democracy in 1991, the country's strongman, President Daniel arap Moi, and his Kenya African National Union (KANU) cynically manipulated extra-state violence to frustrate democracy, secure victory and retain power during the 1992 and 1997 multiparty elections. KANU's electoral victories in 1992 and 1997 were massively assisted by a badly splintered opposition elite and an entrenched culture of impunity relating to the use of informal violence. However, prospects for democracy in Kenya were enhanced by the resounding defeat of KANU by the National Alliance Rainbow Coalition (NARC), itself an alliance of Mwai Kibaki's National Alliance Party of Kenya (NAK) and ex-KANU rebels in the Liberal Democratic Party (LDP) led by Raila Odinga. NARC's presidential candidate. Mwai Kibaki won 62% of the votes in the 27 December 2002 presidential elections against 31% for Moi's handpicked candidate, Uhuru Kenyatta, setting the stage for the 30 December 2002 swearing of Kibaki as Kenya's third president. For the first time since independence, the ruling party was defeated and conceded power. However, Kenya's 2007–2008 crisis marked the weakening of the wave of popular enthusiasm and optimism that had ushered Kibaki to power in 2002.

The fragmentation of the power elite and the electoral crisis

Although earlier multiparty elections in 1992, 1997 and 2002 were fought along ethnic lines in most Kenya's communities, ethnic polarisation reached a fever pitch in the 2007 election. Despite the initial unity and popularity of the NARC-led government, it had soon succumbed to elite fragmentation as a result of endemic friction between the LDP and NAK wings of the party. This friction was intensified by the failure of Kibaki to honour a memorandum of understanding (MoU) between his NAK and Odinga's LDP which had provided for an equal share of cabinet positions and most other high-level appointments, and also for consensus-based decision-making as the basis of NARC's interethnic coalition. Thereafter, the 21 November 2005 Kenyan constitutional referendum irreversibly split the NARC elite right down the middle between those inclusive of Kibaki and many government officials, who backed the government-sponsored constitutional draft, and their rivals in LDP and KANU. After voters rejected the draft constitution by 58.12% to 41.88%, (for the government had unilaterally amended a previous draft drawn up through a broad-based consultation process to increase the powers of the executive), Kibaki threw LDP leaders out of the cabinet in a reshuffle aimed at reasserting his political authority. During the 2005–2007 interregnum, LDP joined KANU to form a loose opposition coalition, the ODM, while NARC politicians loyal to Kibaki formed a new party, NARC-Kenya. However, the ODM itself unravelled ahead of the 2007 election in August 2007 into two distinct parties, the Orange Democratic Movement Party of Kenya led by Raila Odinga and the Orange Democratic Movement-Kenya (ODM-K) led by Kalonzo Musyoka. The faction of KANU led by Uhuru Kenyatta joined the PNU, hastily formed in September 2007 as a vehicle of Kibaki's re-election bid.

The disintegration of the NARC elite occurred against the backdrop of intense populism based on a resurgence of ethnic nationalism which underpinned the campaign and voting during the elections and prepared the ground for the post-election violence. During the campaign, ODM built on the anti-Kikuyu rhetoric that had ensured its victory in the November 2005 referendum. As Chege rightly observes, the party's grassroots campaign turned the election into a contest of ‘forty-one tribes against one’ and ‘Kenya against the Kikuyu’, with the campaign highlighting Kikuyu domination of government and the commanding heights of the economy, including banking and trade, and blaming Kikuyu success for the marginalisation suffered by the other groups (Chege 2008, 133).

Besides ethnicity, however, there were also aspects of the 2007 election that were distinctive. Cheeseman (2008) identifies four of these features. First, Odinga's signing of a Memorandum of Understanding with representatives of the Muslim community for the first time in Kenyan elections turned religion into a mobilising issue, manipulated by both the ODM and PNU in their attempts to win the hearts and minds of both Christians and Muslims. Second, with almost 60% of Kenya's population of some 36 million aged 15 to 35, demographic growth of the youth stratum expanded the electoral register by 6 million to nearly 14 million voters. The increase in the proportion of younger voters turned generational competition into an important axis in electoral politics, with many younger voters tending to support the ODM, which campaigned under the banner of ‘change.’ Despite this, politics continued to revolve around ethnicity. Third, the head-to-head competition between the two main presidential candidates combined with the recent rise of an independent polling industry saw opinion polls becoming a contested feature of the campaign, with politicians challenging or endorsing polls according to whether or not they were in their favour. However, as Tom Wolf (in this collection) rightly points out, with opinion polls using different sampling frames, variant polling results may have contributed to political instability. Finally, the 2007 contest was Kenya's first election to tap into the information age, with the spread of communication technology and the youthful electorate ensuring that the battle for votes, and then lives, was fought in cyberspace through the internet, emails and especially SMS text messages. While undercutting the capacity of the state to control information, the new technology facilitated the rapid spread of rumour as much as ‘truth’. ‘Never before had Kenyan voters been privy to so much election-related information, and never had it been harder to sort facts from fiction’ (Cheeseman 2008, 169). Thus, following the election, these new media were as full of messages of hate as of reconciliation, and mobile phones were used to mobilise both violence and coordinate humanitarian relief. In an uncanny way, these new factors tended to augment rather than neutralise the divisive and negative force of ethnicity that ended in turning Kenya's election dispute into an intractable conflict.

Election as war: The 2007 election crisis

Although the period preceding the election saw escalation of violence in many parts of Kenya, (tensions were high, with some 600 people being killed in political violence in the three months before polling day), there is consensus among many observers that the voting itself was relatively peaceful, giving rise to optimistic predictions that Kenya would witness yet again a peaceful change of guard. As the initial results started trickling in mainly from the ODM strongholds, Odinga registered a strong lead in vote counting on 28 December. But trouble began on 29 December when the lead shrank to only 38,000 votes with almost 90% of the votes counted (180 out of 210 constituencies), with most of the remaining votes in Kibaki's strongholds. Hell broke loose on 30 December when the ECK announced Kibaki the winner with a popular vote of 4,584,721 (47%) over Raila Odinga‘s 4,352,993 (44%) and Kalonzo Musyoka's 879,903 (9%). In parliamentary results, ODM won 99 seats over 43 for PNU, but counting allied parties, ODM had only 103 parliamentarians to PNU's 104 (ECK 2007).

Within minutes of the announcement of Kibaki's victory, he was sworn in for a second five- year term at a hastily organised ceremony at State House before a handful of guests (but excluding diplomats) and the Kenya Broadcasting Corporation, despite calls for a recount by the opposition and international observers. This fostered a widespread perception that the count of the presidential election was modified in favour of Kibaki. In the ODM strongholds, supporters of Raila Odinga, who felt that their candidate had been cheated of victory, erupted in violence and protest demonstrations that soon degenerated into rape, looting, and indiscriminate killings. Two immediate factors contributed to the election dispute degenerating into violent conflict. The first factor is the manner in which the ECK mishandled the vote counting. The chair of the Electoral Commission, Samuel Kivuitu, conceded that irregularities had occurred. Further, Kivuitu also admitted that there were some problems in the vote counting, with some constituencies reporting a turnout rate way above 100%. Yet most inflammatory of all was Kivuiti's damning admission shortly after the proclamation of Kibaki as president that he had been subject to undue pressure, and that he could not say with certainty that Kibaki had actually won the poll. As a result, international observers, including the European Union, declared the election ‘flawed’, blaming the ECK for its failure to establish ‘the credibility of the tallying process to the satisfaction of all parties and candidates’ (Reuters, 30 December 2008). The second factor is the absence of a credible mechanism for resolving electoral disputes. Although the ECK insisted that electoral disputes were a matter for the courts, ODM refused to take the matter to the courts, pointing out that they were controlled by Kibaki, who had nominated six judges, two to the Court of Appeal and four to the High Court, a few days to the election (Ndegwa 2007).

The roots of the crisis

The causes of the post-2007 election violence are complex, multiple and interrelated. Jeffrey Gettleman, writing in the New York Times, claimed that Kenya's electoral violence had ‘tapped into an atavistic vein of tribal tension that always lay beneath the surface in Kenya but until now had not produced widespread mayhem’, feeding a widespread view of an inherent propensity for fratricide, cynicism, and economic nihilism within African societies (Gettleman 2007). However, the broad argument that runs through this collection is that the starting point of understanding lies in a comprehensive grasp of Kenya's political economy, especially historical injustices relating to the land question in the colonial and successor post-colonial states. Yet we need to begin the analysis by sketching out some key outlines of Kenyan post-colonial political economy, which provided the long-term causes of the 2007–2008 crisis.

Economy, population growth and poverty

NARC came to power in 2003 on the platform of jump-starting the economy, creating 500,000 jobs a year, ending corruption, improving public services, and fast-tracking constitutional reforms to devolve power and decision making from the imperial presidency to the grassroots. Kibaki's first administration (2003–2007) witnessed a stunning success in economic recovery as growth rose from 3.4% in 2003 to some 7% in 2007. National poverty levels fell from an estimated 56% in 1997 to 46% in 2006 and per capita incomes rose for the first time since the 1980s. NARC delivered on a promise of free primary education for all children. The government fostered macroeconomic discipline, reduced deficit spending, introduced stricter enforcement of regulations which doubled tax revenue and improved the business environment, privatised failing state-operated enterprises, and attracted substantial new foreign investment and the flow of international aid from traditional Western sources and new Eastern sources, especially from China.

However, this economic record was not matched by an equally robust political stewardship. Although the government erected the legal and institutional framework to combat corruption, its record of prosecuting and convicting high-level corruption before and after 2003 was dismal. In 2004 the government set up a judicial commission to investigate the Goldenberg scandal, a US$800 million Moi-era rip-off involving government rebates for fake diamond exports, and released the findings of the commission in February 2006. But top leaders in both the government and in the opposition who were implicated went scot-free. Meanwhile, the government generated mega-scandals of its own. In early 2005, some of its officers were allegedly involved in the Anglo leasing scandal, a series of security contracts with official payoffs. A raid of the Standard newspaper and its television station by the police in March 2006 reflected a return to Moi-era authoritarian undertows. Further, the NARC government not only failed to stamp out the Moi-era culture of impunity relating to mega-corruption but also extra-state violence. Worse, the Kibaki inner circle stymied the process of constitutional reform promised ahead of the 2002 election to reverse a curtailment of presidential powers proposed by a constitutional convention, thereby alienating the Kenyan public and civil society (Whitaker and Giersch 2009).

The 2007 election took place against the background of a meteoric growth in the Kenyan population, deepening poverty, widening inequality and youth underemployment. Kenya's population, which occupies a total of 582,650 square kilometres (224,962 square miles), an area twice the size of Nevada and slightly larger than France, had jumped from 8.5 million at independence in 1963 (Leys 1975, 40) to an estimated 38 million in 2008 (Central Intelligence Agency July 2008). Competition for land among Kenya's 42 ethnic groups and the small white and Asian populations is accentuated by the fact that over 80% of the land is either arid or semi-arid and over 75% of the population is concentrated in the remaining high potential agricultural belt that runs north-west from the capital, Nairobi, to the Ugandan border. Over-reliance on land is reinforced by around 80% of the workforce being engaged in agriculture, with diminishing land and opportunity in rural areas engendering migration to urban areas in search of jobs and a better standard of living by the largely unemployed young people, who comprise 60% of the population (Kenya National Bureau of Statistics 2007). Although Kenya has a far more advanced industrial sector than its neighbours, its weak manufacturing, trade, finance, insurance and real estate sectors (which account for little more than 20% of gross domestic product and provide less than 300,000 jobs in the formal sector) have increased pressure on the government itself to provide jobs (Republic of Kenya 2003; World Resources Institute 2007). The most obvious point to be drawn from these basic figures is that as Kenya geared toward the 2007 election, demographic growth had already placed the country under increasing stress.

Confronted with acute poverty, inequalities and unemployment in the context of ethnic polarisation and the resurgence of ethno-nationalism, Kenya's fragmented power elite resorted to populism and manipulation of genuine economic grievances and disaffection to win the vote of the poor. These economic realities produced two diametrically opposed campaign strategies by the power elite. For its part, the PNU hurriedly cobbled together as a vehicle for Kibaki's re-election, took as its anthem and slogan ‘Kibaki aendelee’ (Let Kibaki continue) as the basis of its campaign, which looked to an assumed trust among Kenyans in Kibaki's record of economic delivery and competent stewardship. Trumpeting this record, the party's pundits asserted that a Kibaki government was better placed to improve the quality of life for the Kenyan people, as opposed to ODM's ‘empty rhetoric’ (domo). But, despite its impressive economic performance, the PNU failed to counter effectively the appeal of the ODM mantra of ‘change’, which was aided by the galvanising global campaigns in Eastern Europe and Obama's America. The ODM had promised its youthful constituency a new world with free schooling, jobs, cash for the poor, roads and hospitals. Indeed, Odinga, in his September 2007 speech accepting the nomination as ODM's presidential standard-bearer, stated that ODM would end Kenya's ‘economic apartheid’ under which one black group (Kikuyu) had all the privileges. This formed the basis of ethnic populism underpinned by the ‘forty-one tribes against one’ campaign slogan. In this regard, the Kibaki government and the Kikuyu appeared in the campaign as one, and were collectively accused of perpetuating inequalities in ‘national resource distribution’ allegedly harking back to the Kenyatta years.

Ethnic populism: ‘The forty-one-against one’ rhetoric

Analysts of the post-2007 crisis have focused attention on the role of ethnicity in shaping the character of Kenya's political parties and its party system in the multiparty era. In this regard, Sebastian Elischer, drawing on a constructivist perspective of ethnicity inspired by Donald Horowitz's works, has identified three party types in Kenyan politics: the mono-ethnic party, the multi-ethnic alliance type and the multi-ethnic integrative type. He eloquently concludes that even as ‘Kenyan parties have increasingly incorporated diverse communities, they have consistently failed to bridge the country's dominant ethnic cleavages’ (Elischer 2008, 1). Both PNU and ODM were multi-ethnic coalitions forged to win power.

The ubiquity of multi-ethnic party coalitions reflect the demographic reality that Kenya is a country of ethnic minorities as opposed to the clear ethnic bifurcation in countries like Rwanda and Burundi. Although the Kikuyu, with their ancestral homeland in Central and Eastern provinces, constitute the largest of the groups, they are only about 22% of Kenya's nearly 38 million people. Next to the Kikuyu are the Luhya (14%), located mainly in Western Province; the Luo (13%), who mainly inhabit the Nyanza Province; and the Kalenjin (12%), who mainly reside in the central and northern parts of the Rift Valley, the Kisii (6%) in Nyanza and the Meru (6%) in Eastern Province (Central Intelligence Agency, 24 July 2008). Since colonial days, political tribalism and inter-ethnic struggle for the control over the state have revolved around the Kikuyu, Luhya, Luo and the Kalenjin groups, who occupy the high potential agricultural belt from Nairobi and its environs to Lake Victoria and the border with Uganda. Because of the overwhelming dominance of these four political tribes at the national political stage, the vast majority of the other 38 Kenyan groups and racial minorities oscillate between ‘moral ethnicity’ and political tribalism, especially when they are mobilised into one of the political alliances of the four ‘political tribes’ during elections.

In a divided society of ethnic minorities, Kenyan politics has, predictably, revolved around ethnic coalitions to win competitive national elections. As a result, the imperative of forming competitive ethnic coalitions has produced two ethnic dynamics. The first is to unify and form large ethnic conglomerations often from sub-tribes or neighbouring ethnic cousins. Indeed, the Luhya group is a conglomeration of 18 sub-tribes (Bukusu, Maragoli, Banyala, Banyore, Batsotso, Gisu, Idakho, Isukha, Kabras, Khayo, Kisa, Marachi, Marama, Masaaba, Samia, Tachoni, Tiriki and Wanga). Similarly, the Kalenjin emerged in the 1950s as a group comprising nine sub-groups (Elgeyo, Kipsigis, Marakwet, Nandi, Pokot, Sabaot, Terik, Tugen and Sebei). With the onset of political pluralism in the early 1990s, the Kalenjin elite mobilised other ‘pastoral’ groups in the Rift Valley into a mega-tribe known as KAMATUSA (an acronym for Kalenjin, Maasai, Turkana and Samburu) to neutralise the coalition of the Kikuyu, Luhya and Luo groups in the new opposition party, the Forum for the Restoration of Democracy (FORD).

However, the formation of ethnic conglomerations for political mobilisation has not been limited to smaller groups. Following the collapse of the Kikuyu-Luo détente that swept KANU to victory in the 1963 elections, the Mount Kenya elite formed a new regional grouping called GEMA, an acronym of the Kikuyu and their Embu (1.2%) and Meru (6%) cousins, far and away the largest single identity bloc in Kenya consisting of nearly 30% of the total Kenyan population. Their Kamba kith and kin, who make up 11.4% of the Kenyan population, have historically gravitated towards the GEMA bloc, except in 2007 when Kalonzo Musyoka's ODM-K drew the Kamba vote, although he joined the PNU alliance in January 2008, giving the PNU alliance a post-election popular vote of 56%. Further, the Kisii, a bantu group in Nyanza which comprises 6% of the Kenyan people, has also tended to align itself politically with the Mount Kenya groups. Combined with the prominent role of the Mount Kenya groups in the commanding heights of the Kenyan economy, the ability of the Kikuyu to tap a vote-rich coalition has transformed the group into a perceptual majority that has always been a target of fear and counter-mobilisation by other groups. An outcome has been that Kenya's ethnic-based post-colonial politics has given impetus to a phenomenon of blaming the ills of society on the Kikuyu, which Kenyan historian Macharia Munene (1998) traces back to colonial invention and manipulation of political ethnicity. It was this that underpinned ODM's ‘forty-one-against one’ campaign slogan.

The second trend in ethnic formation is the formation of multi-ethnic alliances that have revolved around one or more of the four ‘political tribes’, the Kikuyu, Luhya, Luo and Kalenjin. At any one time, the winning coalition is the one that is able to win the support of the majority from the four groups, which has implications for the stability of Kenya after an election. As Table 1 shows, each of the three Kenyan presidents, Jomo Kenyatta, Daniel Moi and Mwai Kibaki, ascended to power through a strong coalition of two or more of the four ‘political tribes’. This has generated a sort of formulaic approach to politics. Once in power, the success or failure of the incumbent president, and indeed the legitimacy and stability of their governments, have largely been measured against their ability to cultivate and maintain the support of a broad, multi-ethnic coalition of the four major groups.

Table 1. Elections and multi-ethnic party coalitions, 1963–2007.

President	Election	Party and main ethnic coalitions	Impact on democracy/stability
Jomo Kenyatta	1963	KANU (Kikuyu and Luo)	Victory, relative stability
	1969	(Kikuyu, Kalenjin)	Authoritarianism,
	1974		relative stability
Daniel arap Moi	1979	KANU (Kalenjin, Kikuyu, Luhya)	Legitimacy and relative stability
	1983	KANU (Kalenjin)	Authoritarianism
	1988		and resistance
	1992	KANU (KAMUTUSA)	Post-election violence and instability
	1997	KANU (KAMATUSA)	Election violence and instability.
Mwai Kibaki	2002	NARC (Kikuyu, Luo, Luhya)	Victory over KANU and post-election stability
	2007	PNU (Kikuyu/Embu Meru, Luhya (Bukusu), Kisii)	Post-election violence and instability

In this regard, the scramble for the support of the main ethnic groups between the PNU and ODM turned the 2007 election into a high-risk election, and the most narrowly contested in Kenya. For its part, the ODM built on the anti-Kikuyu rhetoric of the 2005 referendum, turning its campaign into a contest of ‘forty-one tribes against one’ and ‘Kenya against the Kikuyu’ and highlighting Kikuyu domination of key economic areas, not only trade, banking, government, commercial farming but also, critically, land ownership. The Kikuyu were blamed for contributing to the marginalisation of other groups. As a result, the ODM managed to forge the alliance involving the Luo, Kalenjin and part of the Luhya against the Kikuyu allied mainly to their Mount Kenya cousins (Meru, Embu and Mbeere), the Kisii and part of the Luhya. This ethnic alignment was responsible for the closely contested election. But ethnic disputes and grievances over land and citizenship became key issues in the campaign, which largely contributed to the post-election violence and displacement.

Land, citizenship and displacement

The existence of historical injustices relating to access, ownership and use of land traceable to the legacy of colonial land alienation and dispossession of Africans provided fodder to the political elite during the 2007 election campaigns. Rutten and Owuor, Kanyinga and Kamungi (in this collection) make a strong case that the land issue constitutes the major structural factor underlying ethnically driven electoral and political violence in Kenya. Colonialism alienated and gave Kenya's best land to a small settler population concentrated in the ‘White Highlands’ and restricted the mass of dispossessed African population to tribally-defined ‘native areas.’ Colonial injustices against the Africans relating to land sparked the violence that characterised the anti-colonial war by the ‘Land and Freedom Army’ (or Mau Mau) fighters. Far from addressing these deeply entrenched injustices, reallocation of land in the late colonial and post-colonial period produced highly skewed and unequal patterns of land access and ownership along class and ethnic lines. Kanyinga argues that the post-colonial land settlement not only entrenched deeply felt resentments between ethnic groups but built upon them by widening divisions between landed and the landless, the rich and the poor.

Inevitably, these dynamics have turned the land question in Kenya into a highly charged issue that has shaken the foundations of the post-colonial ‘nation-building’ project. As Karuti Kanyinga argues below: ‘In Kenya, the land question appears to be the fulcrum around which major political events revolve’. In short, the immediate availability of land-based ethnic grievances and tensions for exploitation by the elite for political ends not only provided the backdrop to much of the violence that threatened to tear the country apart in the wake of the 2007 election, but also defined the major fault lines of the violence and the displacement that occurred.

The contributions by Prisca Kamungi, Karuti Kanyinga, and Marcel Rutten and Sam Owuor in the present collection all provide rich accounts, albeit with different emphases, of how competition for land expressed in ethnic terms resulted in political violence after the 2007 elections. However, the land question, which formed the central axis of the 2007 campaign, was viewed through the prism of the more serious problem of regional economic inequality along ethnic lines. Cast within the larger issue of ethnic populism, the inequality discourse resurrected the controversial ethnic-based form of federalism known as majimboism that Kenya had abandoned in the 1960s (Anderson 2005, 547–64). Re-emboldened advocates of majimboism now prescribed the system as the logical path to regional economic equality and the cure for the problem of marginalisation of other less-developed regions (Muluka 2006a, 2006b). Following the worldwide trend where political campaigns have targeted ‘economically dominant minorities’ to win closely contested elections, with dire results for global stability (Chua 2003), crusaders for majimboism ‘stoked ethnic tensions with an effective campaign of disinformation in the run up to the elections, pitting the Kikuyu against Kenya's other African ethnic groups’ (Chege 2008, 133).

In essence, the majimbo crusaders attributed the Kikuyu economic edge over other Kenyan groups to the political patronage they had enjoyed under Kenyatta's presidency, accusing Kenya's first president of ‘pouring’ the Kikuyu from Central Province into the pristine homelands of the pastoral Kalenjin and Maasai in the Rift Valley. After independence, the Kikuyu obtained preferential treatment in government-financed settlement schemes for previously white-owned farms in the Rift Valley. But as Tabitha Kanogo has shown, long before Kenyatta assumed power in 1963, in response to British expropriation of their lands, the Kikuyu had already moved to parts of the Rift Valley by 1903 where their numbers rose to an estimated 150,000 by the 1930s and constituted 26% of the total African population region, according to the 1948 census (Kanogo 1987). As Gisemba (2008) has shown, together with other the Kenyan groups, mainly the Luhya, Kisii, Kamba, Luo and Kalenjin, the Kikuyu acquired land in the Rift Valley after independence in three ways: British and World Bank settlement programmes; state-financed transfers of British-owned land to groups of Kenyan buyers across ethnic lines; and market transactions. Moreover, under Kenyatta, poor Kikuyu became victims of expropriation of their land in central Kenya to create room for the large estates of the propertied Kikuyu class and ‘home guards’ that had aided the British counter-insurgency against the Mau Mau, creating serious social cleavages and tension in the Kikuyu homeland. Ultimately this reinvention of history relating to migrations and acquisition of land in the Rift Valley provided the ideological fire for the opposition's ‘forty-one-against-one’, creating the impression that ‘the Kikuyu had set themselves up for resentment and retaliation’ in the 2008 post-election violence (Chege 2008, 134). In the end, majimboism, which thrived on the divisive binaries of ‘natives’ and ‘settlers’, ‘indigenous’ and ‘foreigners’ elevated notions of ethnic citizenship over civic citizenship based on the civic nation and stripped those condemned as ‘foreigners’ of their rights, thus setting them up for violent expulsion and displacement from their land, which Prisca Kamungi eloquently analyses below.

Nevertheless, the history of violence and displacement in multiparty Kenya runs deeper than the 2007 election. Rutten and Owuor (2009) cite serious clashes between Maasai and Kikuyu immigrants in Narok in 1992 and the August 1997 attacks by Coastal youths against non-indigenous residents in Likoni in Mombasa district (this leaving as many as 100 dead), which displaced an estimated 100,000 people. In these and other cases, politicians played upon local resentments against exclusion from land, water rights, employment opportunities and access to other resources to provoke violence against outsiders, not least because the presence of the latter within their constituencies presented a numerical threat to their winning the forthcoming election. Furthermore, as has been widely pointed out and is discussed below, the inheritance of the Westminster-style, first-past-the-post election system from Britain encourages a zero-sum style of politics which in the Kenyan situation lends itself to ethnic campaigning or, in a more extreme variant, to ethnic cleansing. Yet it is important to highlight that, in contrast to 1992–1993 and 1997–1998, in 2007 violence between the Masai and the Kikuyu was virtually absent. Those interviewed by Kagwanja for this article indicated that earlier waves of violence and displacement of the Kikuyu had taken a heavy toll on the local economies in Masai land, giving rise to fear that renewed violence would push the economy further to the ropes. In line with this, and as Kamungi argues, the danger of violence diminishes when politicians stress the common interests shared by ‘insiders’ and ‘outsiders’. Thus it was that there was no repeat of the violence that had afflicted the Rift Valley in 1992 and 1997 during either the 2002 election or the 2005 referendum campaigns, (although as Kamungi notes, as many as 352 people were killed during party constituency nomination campaigns, mostly within NARC, prior to the election in 1992).

In a word, displacement as an integral feature of the Kenya crisis was inextricably linked to land issues, rendered more urgent by dramatic population growth and the resurgence of ethnicity overlaid by class and political manipulation. But land had become almost inextricably entangled with other more specifically systemic and political processes: the post-independence construction of an ethnically-based patronage politics centred around single-member electoral constituencies; the centralisation of power under the presidency; and an erosion of state capacities, notably through the growth of armed militias associated with politicians and unprecedented diffusion of violence during the 2007–2008 crisis.

Weak institutions and the diffusion of violence

Susanne Mueller has provided a compelling analysis, arguing that Kenya was poised to implode even before the election dispute provided the spark for the violence in early 2008. In analysing why the violence in 2007 reduced Kenya to near disaster, she identifies three underlying factors as follows:

A gradual decline in the state's monopoly of legitimate force and a consequent generalised level of violence not always within its control; deliberately weak institutions, mostly overridden by a highly personalised and centralised presidency, that could and did not exercise the autonomy or checks and balances normally associated with democracies; and political parties that were not programmatic, were driven by ethnic clientism, and had a winner-take-all view of political power and its associated economy byproducts. The argument here is that: violence was diffused, could be ignited easily, but not controlled, and was not; institutions outside the presidency normally associated with vetting a contested election were not viewed as being sufficiently neutral to do so and did not; and the nature of Kenya politics predisposed both leaders and followers to seek politics as a do or die zero-sum game, which is what the elections became. Had the election not been so close, these same factors may have been held in check for a while. Nevertheless, they were dangerous and looming problems. (Mueller 2008, 186)

Her argument is sustained by the suggestion that false optimism among scholars and policy-makers about Kenya's democratic transition prior to the 2007 election emanated from their focus upon the formal aspects of institutions such as parties, parliament and administrative structures rather than upon the incentive systems that guide the political behaviour of actors within them and dictate how institutions ‘really operate’. In particular, she stresses the importance of assessing the role played by violence in electoral competition, of which it may itself be a product and hence not necessarily easily addressed by institutional reform, and the general lack of accountability for a wide range of de jure violations (ibid.).

Mueller argues that while the post-independence period under President Kenyatta saw violence used to repress opposition politics, it was largely dispensed and controlled by the state. However, the succeeding Moi presidency was far more repressive because he had considerably reduced opportunities for the dispensation of patronage. Kenyatta had used opportunity provided by the decolonisation process to reward followers with positions in the civil service and with the farms of departing settlers, and was buoyed by expansive economic conditions. In contrast, Moi faced a recalcitrant Kikuyu political elite, multiple demands from his own Kalenjin constituency, and a growing population waiting for perks while confronted by a collapse in coffee prices and a rise in the cost of oil. Whereas Kenyatta ‘could give without taking away, Moi had to take away before he could give’ (Mueller 2008, 188), although, because his power base was in poorer parts of the country, he was inclined to reward them at the cost of the more productive areas.

This resulted in the measures he used to consolidate support and control opposition being much cruder and more repressive, involving notably the diversion of state resources to enrich his followers, assaults upon the economic foundations of the Kikuyu elite in agriculture, and an increased use of state violence against his critics and opponents.

Critically, too, Moi and his allies resorted to extra-state violence in the form of vigilantes linked to the ruling KANU operating side by side with the official security forces (Katumanga 2005, 505–20). The use of informal violence increased notably after Moi's reluctant concession to return Kenya to multipartyism in 1991. In the subsequent multiparty elections in 1992 and 1997, KANU politicians hired and used gangs to kill and displace opponents in key electoral areas, a development which encouraged a wider ‘privatisation of public violence’ and led during the 1990s to the ethnicised clashes in, notably, the Rift Valley, Western and Coast provinces (Kirschke 2000; Klopp 2001). Further, squandering its own monopoly of force, the increasingly criminalised state facilitated an alarming growth of urban gangs which became available for hire, operated protection rackets, and generally preyed on the poor and endangered public order (Anderson 2002, 231–55; Kagwanja 2001, 275–95). Gangs like the Mungiki were at the disposal of politicians who eagerly used them to keep themselves in power (Kagwanja 2006, 51–75).

The wrangling NARC leadership that took over from Moi failed tragically to demolish the architecture of informal violence that emerged in the latter years of Moi's long tenure. As Kagwanja (in this collection) demonstrates, this extra-state violence was quickly diffused after chaos erupted on 31 December 2007. The diffusion of violence manifested itself in three inter-related categories which emerged, mutated and defined Kenya's post-election impasse, including the spontaneous violence in ethnically mixed urban areas, especially Nairobi and Kisumu; organised violence by the power elite in the Rift Valley region, followed by largely organised ‘retaliatory violence’ by criminal gangs such as the Mungiki and the Taliban; and state violence by security agencies, widely accused of indiscriminate killings, rapes and violation of human rights. The diffused violence overwhelmed the state, eroding its capacity to function and protect citizens during the 2008 crisis.

The normalisation of violence helped to undermine state institutions, the constitution and the law. To be sure, over time the formal rules of the political game were changed (by 1991 the constitution had been amended some 32 times) to buttress the executive. However, informal norms undercut formal rules in favour of a personalised presidency, as Moi erased distinctions between party and state, dissolved checks and balances, created a culture of fear, rendered the judiciary servile, visited torture and violence upon those who did not conform, and looted the state to feed his supporters, thereby entrenching corruption as integral to the state's functioning. The crisis after the 2007 election was, in essence, a reflection of failed attempts by the power elite to agree on how to devolve power from the imperial presidency to a reformed, but integrated executive. This poignant point was made by the Kenyan Commission of Inquiry (the Waki Commission) as follows:

The attempt to reduce the personal power [of the presidency] that had been accumulated by former President Moi initially was the reason opposition forces sought to introduce the post of prime minister. This culminated in an informal Memorandum of Understanding (MoU) before the 2002 election between the then opposition coalition under which the coalition agreed to introduce the post of prime minister after the election. Once elected, however, President Kibaki reneged on the MoU. Discussions continued concerning constitutional change and the devolution of power. The Kibaki government then came up with a draft constitution put forward by Attorney-General Amos Wako watering down some of the provisions in the draft agreed to during the ‘Bomas’ discussions. The Wako draft was put to the public at a referendum in 2005, where voters rejected it. […] As soon as the MoU was scuttled, a group led by Raila Odinga left the NARC coalition government. President's Kibaki government was perceived as being unwilling to abide by its pre-election agreement with its partners and as retreating into an ethnic enclave. This was criticised by the public and was seen as an attempt by the so-called Mount Kenya Mafia to keep power to itself rather than share it. Even though the MoU was not a legal agreement, the Kibaki government's turning away from it and removing from government the group of ministers associated to Odinga had the effect of increasing the polarisation of politics along ethnic lines. Even though the 2005 referendum was peaceful and the results were accepted rather than contested, the parameters were nevertheless drawn. With the ethnic political fault lines clearly drawn after 2005, and the need to win the presidency seen as paramount, tensions began to mount. (Republic of Kenya 2008, 30)

This phenomenon of the deliberate weakening of state institutions was to prove a major precipitating factor in the electoral violence of 2007.

The third precipitating factor made suddenly worse by the 2007 election was the non-programmatic nature of clientist parties based on ethnicity. Kenyan political parties, argues Mueller, have ideologies, policies and programmes that are largely indistinguishable, with politics being viewed as a winner-takes-all zero-sum ethnic game. The objective is therefore to secure control of the state in order ‘to eat’, and to ensure that a co-(or allied) ethnic becomes president, this requiring the formation of alliances and cross-ethnic coalitions. Politicians thus obtain power by using ethnic arithmetic and clientage as mobilising factors, the issue simply being how to win power in order to access state resources (although, as Wolf observes below, because ethnic alliances have certain policy implications, as with the differences between the ODM and PNU over ‘majimbo’, the ethnic arithmetic can be more complicated than Mueller seems to imply). As a result, parties move in and out of alliances, and politicians shift parties without compunction according to calculations of advantage, so that – regardless of party – most of the political elite who confronted each other in 2007 had at one time or another served in government together. Most notably, Kibaki had served as Moi's vice-president from 1978 to 1988 before defeating the latter's nominee in the 2002 election; and Odinga had likewise served as secretary-general of KANU and cabinet minister in Moi's government towards the end of his long incumbency, before backing Kibaki for president and then serving in his government before breaking with him over the constitutional referendum in 2005. The lengths politicians are prepared to go to get their leader into the presidency and themselves into government predisposes them to use violence to achieve their ends, and helps explain the ease with which violence was mobilised to contest the result of the 2007 elections (Lafargue 2008; Lafargue and Katumanga 2008).

Mediation, power sharing and paralysis

The intensity and devastation of the 2007–2008 violence shocked the Kenyan elite, the public and the international community. Within the first week of violence, influential members of the business community (mainly owners of flower farms, hotels and coffee estates who had benefited from the Kibaki administration's economic liberalisation, improved infrastructure, a tourism boom and exponential growth of agro-industry and the Nairobi Stock Exchange), facilitated by the World Bank resident representative, Colin Bruce, initiated a discreet move to broker a political agreement between the ODM and PNU to halt the killings (East African, 14 January 2008; International Crisis Group 2008, 21). Notably, this group of company executives recognised the need for power sharing between Kibaki and Odinga to halt the chaos and stabilise the country.

By the time the chair of the AU, John Kufuor, landed in Nairobi on 8 January 2008, the mediation by the corporate class had already produced a document detailing ‘Principles of Agreement’ to be signed on 10 January 2008 as the basis of a more comprehensive process to address the root causes of the electoral dispute and the consequent violence and arrive at a permanent political settlement. However, the document fell through when it was disowned by the Kibaki government, which insisted on the end to violence before any deal could be discussed (Kagwanja interview with a senior government official, 8 January 2008). Both Gilbert Khadiagala and Monica Juma detail the role of local and regional actors in resolving Kenya's conflict, this leading to the appointment by the AU of a high-powered team of eminent African personalities led by the former UN secretary-general, Kofi Annan, and which included Graça Marcel (wife of Nelson Mandela) and the former president of Tanzania, Benjamin Mkapa. The AU mediation team had critical high-level support from the United States, Britain, the European Union, and the African Union itself in its bid to restore peace in Kenya.

Ultimately, it was the two party leaders who saw sense in compromise and shook off extremists who, for the ODM, were urging a break-up of the country along party lines and, in the case of the PNU, the use of more force to overcome opposition resistance. On 28 February, the parties reached an agreement on a PNU–ODM coalition government with Kibaki as president, Odinga as prime minister, equally shared deputy prime ministers (two) and a cabinet comprising 40 ministers and 50 assistant ministers reflecting Kenya's ethnic diversity. The new government was sworn in on 17 April at State House.

The mediators based their efforts to resolve the Kenyan conflict on the idea of liberal peace, widely used to broker peace agreements to end deadly conflicts across the African continent from South Africa to Sudan, Congo to Côte d'Ivoire. The theory of liberal peace insists on a negotiated resolution of dispute often in divided society through power sharing between former enemies in a ‘grand coalition government’ of all major parties and leaders; sharing of power between ethnic-based factions; and allocation of executive positions so that all major groups are fairly represented. These practices are in line with the ‘power sharing’ or ‘consociational’ model that Arend Lijphart (1985, 1990) prescribed as alternatives to democracy in deeply divided pluralist societies such as post-election Kenya. Missing in the Kenya National Accord, however, are three key elements of the power-sharing model: autonomy and federalism; the mutual veto of the warring parties; and the electoral system anchored on proportional representation (Sisk 1996). But the agenda for the peace accord provided for far-reaching constitutional and institutional reforms, which were expected to address these issues.

The National Accord prevented Kenya from slipping into the category of failed states, and has endured despite predictions of inherent instability. But what exactly prevented Kenya from falling over the edge? To understand how and why Kenya pulled back from the brink, we need to ask how and why the political elites who had so recently been engaged in bitter electoral competition reached a compromise. Following that, we need to concern ourselves about the longer-term prospects for Kenya avoiding politico-ethnic violence in the future, not only at the next election (expected in 2012) but in the future thereafter. There are three aspects to this.

First, although both political elites and their followers had demonstrated willingness to use violence for political ends, and recognised that they would be unlikely to incur state penalty for doing so, they ultimately proved unwilling to take the country into civil war. Although ethnically divided, the political class had many interests in common; although prepared to deploy violence to gain power, their intention was not to destroy the state. Past experience of collaboration across Kikuyu–Kalenjin lines (in particular) indicated that compromise could be achieved. Meanwhile, the broader middle class (inclusive of many ‘mixed marriages’) was badly affected by the violence, and together with the business community and well-off landowners had much to lose. There were moderate forces in both camps willing to work with non-governmental organisations, civil society and the media to bring about an end to the violence.

Second, domestic momentum towards ending the crisis was backed by external intervention, as outlined by Brown (US and EU), Khadiagala (the region), and Juma (the AU and the continent). The bottom line here was that Kenya was too important to be allowed to become another Rwanda.

Thirdly, there is the issue of whether the present coalition and political moment offers a genuine opportunity for significant reform capable of providing for peaceful elections and democratic outcomes. Senior politicians may understand that, without reform, violence may break out again in 2012, but they are unwilling to make the concessions and changes needed to avert it. This begs the question: Can a politics of patrimonialism reform itself? Can a politics revolving around a centralised presidency, manipulation of ethnicity, use of violence and corruption to buy support from sub-patrons (whilst excluding others) and the selective application of the rule of law reform itself?

While the coalition has continued to hold, Kenya is still at risk, and its democracy trapped in limbo. Besides providing for a power-sharing government, the National Accord called for the establishment of three commissions to deal with the larger agenda of promoting reconciliation, healing, restoration and ending impunity on violence. The Independent Review Commission (IREC), headed by the South African judge Johann Kriegler, was established in March 2008 to investigate all aspects of the 2007 presidential elections and make recommendations to improve the electoral process. Further, the Commission of Inquiry into the Post-Election Violence headed by Judge Philip Waki was established to investigate the causes of the violence and the role of state security agencies, and to suggest ways of preventing, controlling, and eradicating similar violence in the future. Also agreed upon by the parties was the appointment of a Truth, Justice and Reconciliation Commission to inquire into historical injustices. However, the implementation of the report and recommendations of both the Kriegler and Waki Commissions is hampered by continued intense political animosity and intermittent squabbling over the undefined powers of the vice-president, the prime minister and the head of the public service. The power tussles are poised to increase ahead of the 2012 elections. In July, the lead mediator, Kofi Annan, handed over to the prosecutor of the International Criminal Court at the Hague, Luis Moreno-Ocampo, an envelope with the names of Kenyan leaders suspected to have masterminded the post-election violence, and has promised to disclose the names and move speedily to prosecute the suspects. This was after the failure by the government and parliament to agree on the formation of a local tribunal to try the suspects as recommended by the Waki Commission. This development caused panic among the power elite across the political divide, and has the potential to derail the coalition.

Conclusion: The road to 2012 – peace or Armageddon?

The 2008 violence shook the foundations of Kenyan society, and has left deep scars in the public psyche. It is now clear that if Kenya is to hold together, its elite must seriously implement the reforms recommended by the mediation team. As Kasfir (2008) has argued recently, politicians are likely to pursue a minimum rather than a maximum reform agenda (suggesting the importance of civil society pressure to push the envelope). However, if such reform is to achieve a functional, non-violent democratic change then three conditions must obtain. First, there has to be court room action against at least some of the perpetrators of violence in 2007–2008 to both assuage public and international opinion and to stem the culture of political impunity and deter a repeat of the violence in future elections. Second, it has to be non-reversible (Kasfir 2008) or it should seek to change political incentives so that resort to ethnic mobilisation and violence costs rather than builds support (Mueller 2008), not least through a change in the electoral system. Finally, there has to be progress towards demilitarisation of political society accompanied by the depoliticisation of the security forces.

However, reform remains a daunting agenda in the context of power sharing. Chances are high that the power-sharing government may change the rules of the game because politicians now have to deal with three audiences, namely, their immediate patronage networks; their partners in the government; and a more alert set of international actors. At present, although government established both the Kriegler Commission to address the root causes of the violence and make recommendations on the election process, and the Waki Commission to deal with the post-election violence, it is yet to appoint a Truth, Justice and Reconciliation Commission to inquire into historical injustices and oversee the reconciliation of communities. The government has also appointed a committee of experts to pursue the constitutional review process, but this has been bogged down by conflict of interests and wrangles among the elite. The slow implementation of the reforms has given rise to the general concern that the grand coalition is merely papering over cracks, and that deep-rooted tensions within Kenyan society will manifest themselves again, perhaps even before the next elections expected in 2012. This has fostered the view of Kenya's democracy being in retreat, and at risk. At the heart of the Kenyan crisis was the elite dispute over power and the foundations of the post-colonial democratic state. This signals the urgent need to renegotiate these foundations and to draw up a new constitution on the basis of widespread mass involvement rather than domination by the elite.

Central to this renegotiation is the need to reassert the basic democratic rights of citizenship accessible to all Kenyans, including the right to change government through free and fair elections; the entitlement of all Kenyans to own property and vote as they wish anywhere in the country; and the outlawing of hate speech as a political campaign tool. The future of democracy and stability of Kenya depends wholly on carefully implementing the reform to foster unity and reconciliation in the country's deeply divided post-election society.