This chapter provides subnational evidence from Kenya’s Rift Valley and Coast Provinces to show how unstable parties have incentivized elites to organize and sponsor party violence in these places. It also incorporates additional subnational variables, including information on candidates’ anxieties over seats, demographic data, and fine-grained information on grievances to explain where, when, and how violence has been organized in the Rift Valley and Coast.

This chapter introduces the phenomenon of party violence, discusses the scope conditions and central arguments of the book, and offers a methodological justification for the distinct cross-regional comparison of Kenya and India. It also details the multiple data sources used to develop the book’s main claims as well as the subnational research sites investigated in both countries. Substantively, the chapter holds that party instability is an underappreciated factor in the broader instrumentalist literature on elites’ decision-making about conflict. It argues that instability matters because it can make the deployment of violence less costly and risky for politicians and thereby incentivize the production of recurring and severe conflict.

This chapter details the book’s theoretical model, focusing first on elites’ decisions and then on voters’ reactions. It highlights how expected party lifespan stands to impact leaders’ decision-making about violence by shortening or lengthening their time horizons. Politicians operating with truncated time horizons will display a higher propensity for organizing or sponsoring party conflict than their counterparts with lengthy time horizons. The chapter thus holds that the effect of party instability on elite choice is conditioning rather than determinative. While unstable parties do not cause violence, they can incentivize elites to engineer or sponsor violence in certain contexts.

This chapter combines national-level violence and volatility data with in-depth elite interviews to demonstrate the relationship between short projected party lifespans and recurring bouts of ethnic party violence in multiparty Kenya. The chapter proceeds in three phases from the KANU era to the period after the promulgation of the country’s new constitution in 2010. The central findings reveal that although Kenyan voters are not lacking in information about the political nature of party conflicts and actually reject violence-wielding politicians, high levels of party replacement and attendant changes in coalitional arrangements tend to prevent them from holding these leaders to account. As a result, politicians from different parties have been able to organize and sponsor violence on a repeated basis.

This chapter illustrates the relationship between politicians, parties, and communal conflict in India from the 1950s through the late 1980s. Combining national-level violence and volatility data with in-depth qualitative interviews, it shows that the weakening and decline of the Indian National Congress (INC) in the late 1970s spurred an escalation of riot violence across many parts of the country through the 1980s. Since then, however, severe riots have dramatically declined in India, as party stabilization has rendered the risks of provoking such violence prohibitive for many political parties. However, other forms of conflict – including rural clashes and targeted low-level attacks against Muslims – have escalated in recent years under the Bharatiya Janata Party (BJP). The chapter suggests that these newer modalities of conflict are part of the same recalibrated elite strategies that have contributed to declines in communal riots across India.

This chapter offers a subnational accounting of patterns of riot violence in Hyderabad in Andhra Pradesh and Meerut in Uttar Pradesh. It shows that much like at the national level, these cities fell prey to repeated and severe riots when soaring party instability incentivized conflict on the part of both Congress elites as well as politicians from its emerging electoral rivals. However, following the restoration of relative party stability in the late 1980s, both Hyderabad and Meerut have witnessed communal quiescence. The chapter further shows that this quiescence is due to the fact that elites are keen to avoid sanctioning from voters for engaging in conflict.