Elections in Kenya paint a picture of a country severely fractured by ethnic cleavages. The conflicts following the disputed December 2007 presidential election in many respects mirror the character of violence that preceded the 1992 and 1997 General Elections.
The cycle of violence around elections in Kenya appears to support views expressed by President Daniel Arap Moi, who argued that Kenya was best suited for a single party system of governance because multipartyism would result in increased tribalism.
The merging of various parties to form Jubilee Party is informed by this thinking – that Kenyans are hopelessly ethnic and therefore must be herded into a single monolithic party to save the nation.
Still, the question is: Do Kenyans vote exclusively according to ethnic identities, or do they also make individual and policy considerations?
It would be unwise to jump to the conclusion that Kenyans are not ready for democratic competition. Not only do most Kenyans abhor violence and seek to peacefully participate in the selection of their own leaders, but elections – including the December 2007 contest – reveal that voters consider factors other than ethnicity in deciding how to vote.
That notwithstanding, ethnicity has played a central role in Kenyan politics as evidenced by patterns of political mobilisation, resource allocation and public service appointments, notably to the Cabinet and senior parastatal positions.
The introduction of multiparty politics in 1991 brought ethnic politics to the fore as opposition parties quickly splintered according to ethnic groupings. As a result, the first multiparty election held in 1992 was a contest that largely anchored around ethnic alignments, a pattern repeated in the 1997 General Election.
Nevertheless, the view that voting in Kenya is simply a cultural phenomenon was weakened in the 2002 General Election when a broad coalition of ethnic groups supported Mwai Kibaki.
The comparative literature on mass electoral behaviour can be viewed as a debate on the relative importance of social identities or economic interests.
For advanced democracies, elections tend to take the form of a referendum on the economy, with voters rewarding or punishing incumbent political parties at the ballot box, depending on their past policy performance.
For instance, voting in the August 8 General Election may be determined by the high cost of living (price of basic commodities such as cooking flour), corruption, lack of inclusivity in government and the reluctance of the Jubilee government to fully embrace devolution.
Evidence of this sort of retrospective, interest-based economic voting has also been found in Latin America and other parts of the developing world.
More commonly, however, voters in new democracies and deeply divided societies, such as Kenya, are held to rely on cultural attachments when deciding how to vote.
The term “ethnic census” was coined to describe elections in which racial, linguistic or tribal solidarities so strongly predict voting behaviour that elections are little more than a head count of identity groups – Mutahi Ngunyi’s Tyranny of Numbers.’
However, recent literature on this subject indicates how identities and interests can coexist and reinforce. For instance, in India voters consider the size of the ethnic group that each party represents as a means of calculating the likelihood of gaining access to patronage.
With reference to settings as diverse as Spain, Ecuador and Romania, ethnic groups compete peacefully in elections (rather than resorting to violence) when they perceive opportunities to secure places for their representatives within decision-making institutions, including the Cabinet.
The debate between identities and interests is alive and well in studies of African elections. Ethnicity and linguistic cleavages are important in explaining an individual’s support for parties in power in most, but not all, places. Identity voting is strongest in ethnically fragmented societies, but popular evaluations of government performance in service delivery are also important in influencing voting choices.
Besides, the state of the national economy or future expectations of personal economic well-being have even trumped ethnicity in selected elections in Zambia and Ghana.
In South Africa, we find only weak support for expressive voting based on identity alone, but also no support for policy-based interest voting. Instead, which is insightful, voters use information on the assumed ethnic identities of parties, casting ballots for those they expected to best defend their group interests in a context where others are assumed to vote along identity lines.
It is therefore expected that ethnicity will be an important factor in explaining electoral choices in Kenya come August 8, but only as one among several relevant determinants of partisanship.
Whereas people will vote according to their ethnic origins, they will also care about policy interests such as personal economic well-being, the performance of the economy, and the government’s policy record in select issue areas such as fighting corruption, the state of the economy, exclusivity or lack thereof, past economic injustices (real or perceived) and the cost of living.
To confirm this view, we also discover that ethnic voting contradicts Kenyans’ views of themselves as adherents to a national (Kenyan) identity. Furthermore, the importance of ethnicity seems to vary depending on voter’s self-ascribed identity, with “ethnics” more often employing feelings of group identity and “non-ethnics” more often making rational calculations of personal and group interest.
Besides, claims about the dominance of ethnic voting are usually based on broad generalizations arising from analysis of aggregate national data that are not well suited to revealing voter intentions.
However research on voter attitudes and behaviour at the individual level, makes it possible to test generalizations about the effects of ethnic origins and ethnic identity on voting and to weigh these factors against other expressed and inferred motivations for electoral choices.
When this is done, it dawns that voting behaviour is hardly exclusively ethnic driven.
Edwin Wangoli Wanjawa, teaches Political Sociology at Pwani University
e.wangoli@pu.ac.ke
