Post-election violence in Kenya: leadership legacies, demography and motivations

ABSTRACT

To understand the political geography of Kenya’s 2007–08 post-election violence (PEV), locational (territorial) and relational (ethnic community) theories of how context affects politics are combined. This hybrid conceptualization provides a nuanced illustration of how subnational violence patterns are shaped by Kenyans’ ethnic community-level experiences and by regional demography. Communities whose leaders have held presidential office in the past (prior-incumbents) have skill using government institutions to distribute material benefits to supporters, which increases their motivations for political activity, including violence. Where the size of the two prior-incumbent communities – Kikuyu and Kalenjin – are large, the severity of PEV is elevated. The presence of a community without these experiences controlling the presidency acts as a buffer, reducing the severity of conflict. Uasin Gishu and Trans Nzioa counties are used as illustrative cases highlighting the relationship between locational–relational contexts and PEV. While both counties endured substantial electoral conflict, Trans Nzoia’s comparatively diverse ethnic geography diluted the incentives for deadly PEV. Understanding these multidimensional contextual influences for violence improves the understanding of conflict geography.

KEYWORDS:

• political geography
• political violence
• contextual effects
• elections
• territory
• governance
• conflict
• Kenya

ACKNOWLEDGEMENT

The author owes many thanks to Abel Oyuke and Sam Balongo at the University of Nairobi Institute for Development Studies, as well as other friends in Kenya. Three reviewers and the editor offered helpful suggestions that improved this work.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the author

Notes

1 Providing motivation for this paper, in 2011 a middle-aged Meru man living in Eldoret explained: ‘Kikuyu and Kalenjin have more in common than either has with Luo or Luhya because they have been in power. You can’t give them [Luo] control because they won’t know what to do with it’ (author interview 1).

2 ACLED data for 1997–2006 (preceding PEV) show that TN saw almost 50% more conflict than UG. This is because of fighting along the Uganda border (e.g., Sabaot Land Defense Force), but suggests that the level of existing violence alone is not a strong confounding influence.

3 See http://www.opendata.go.ke/datasets/2005-2006-kihbs-takwimu-ya-umasikini-katika-wilaya. These are district data. Districts preceded counties created in the new constitution and were more numerous, but UG’s and TN’s boundaries are the same.

4 UK Economic and Social Research Council (ESRC) grant number ES/R005753/1, ‘Spatial Dynamics in African Political Economy’ (Catherine Boone, Principle Investigator).

5 A strong police presence in certain Luo strongholds could have prevented rioting and looting from escalating and becoming as deadly as violence in other areas. We cannot know after the fact if this scenario would have played out, but acknowledge this possibility. Among the police there probably was a greater willingness to use deadly violence in the Luo slums of Kisumu because many of the security forces personnel were from other ethnic communities (Okia, 2011). The relationship between policing and ethnic politics is nevertheless quite complex (Ruteere, 2011).

Additional information

Funding

This study was supported by the US National Science Foundation [grant number 1003838] as well as the Social Science Research Council.