PO377 Ethnic Conflict and Political Violence

1. PO377 Ethnic Conflict and Political Violence Week 15: Institutional Design as Conflict Management: Executive Structures and Electoral Systems in Divided Societies

2. Lecture Outline • Introduction: Two Examples (Nigeria and Sri Lanka) • Theorising the Relationship between Institutional Design and Violent Ethnic Conflict • A grievance-based explanation • The intrinsic value of political representation • The first instrumental value of political representation • The second instrumental value of political representation • A grievance-based explanation (summary) • The Impact of Executive Structures on Ethnopolitical (In)Stability • Defining presidential, parliamentary and mixed forms of government • The virtues of presidentialism in divided societies • Do these virtues really hold? • The perils of presidentialism in divided societies • A grievance-based explanation applied to the design of executive structures • Choosing Electoral Systems for Divided Societies • Majoritarian electoral systems • Proportional electoral systems • Majoritarian vs. proportional electoral systems • What about vote-pooling electoral systems? • Summary

3. Part I: Introduction: Two Examples (Nigeria and Sri Lanka)

4. First Example: Nigeria • Africa’s most populous country, composed of more than 250 different ethnic groups, the three largest of which are the Hausa and Fulani (29%), Yoruba (21%), and Igbo (18%). • Former British colony, independent since 1960. • First Republic (1960-66); Second Republic (1979-83); abortive Third Republic (1993); Fourth Republic (since 1999).

5. First Example: Nigeria (2) • First Republic = parliamentary; Second and Fourth Republics = presidential. • Horowitz (1990; 1991): Parliamentary system in the First Republic was a crucial factor in bringing about the Biafra Civil War (1967-70). A separately elected presidency, ideally with an electoral formula guaranteeing exceptionally broad support for presidential candidate, combined with a strict separation of legislative and executive powers, has conflict-reducing effects. • Suberu and Diamond (2002): Presidentialism might be equally risky, as the inevitable zero-sum character of presidential elections has exacerbated ethno-regional anxieties and inter-ethnic tensions in Nigeria.

6. Second Example: Sri Lanka • Sri Lanka’s largest ethnic group is the Sinhalese (74%), followed by Sri Lankan Tamils (13%), Muslims (7.5%), and “Indian” Tamils (5%). (1980s data; hard to be accurate for the present though Sri Lankan Tamil population has decreased.) • Former British colony, independent since 1948. • War between Sri Lankan Tamil insurgents and Sinhalese-dominated authorities 1983-2009.

7. Second Example: Sri Lanka (2) • First-past-the-post electoral system used for all parliamentary elections before 1989; since 1989 ‘PR with a twist’ (voters cast their vote (1) for a party and (2) up to 3 candidates within this party (=preferences); seats allocated according to the d’Hondt formula to all qualified parties and among candidates according to number of preferences they received; party with highest number of votes in a district is entitled to a ‘bonus’ seat). • DeVotta (2005): First-past-the-post system not only disadvantaged ethnic minorities, but also led to culture of ethnic outbidding. • The change to PR system – too little too late or simply ineffective, as the culture of ethnic outbidding continues (ibid.)?

8. Part II: Theorising the Relationship between Institutional Design and Violent Ethnic Conflict [following on from last week]

9. A Grievance-Based Explanation • Recap: Grievance-based explanations of intrastate conflict argue that ethnic violence is an expression of grievances about the non-fulfilment of certain value expectations; ethnic groups who perceive that they cannot get the values they feel entitled to, are expected to develop emotions of anger and resentment which – under the ‘right’ circumstances – can translate into violent action (Harff and Gurr 2003). • Theuerkauf (2010) argues that there are three distinct values of being represented politically (next three slides), which is why political institutions that reduce the chances of political representation particularly for ethnic minorities can be expected to increase the odds of ethnic violence.

10. The Intrinsic Value of Political Representation • Intrinsic values relate to the worth of objects and practices by themselves, i.e. objects and practices are intrinsically valuable if they are considered a relevant good per se (Réaume 2000). • The argument that political representation is intrinsically valuable is based on the normative assumption that having a ‘voice’ in the political decision-making process is an intrinsically rewarding experience for all ethnic groups: ethnic groups want to make their interests heard and government to respond to them, as it affirms their status as recognised members of a political community (e.g. Kymlicka and Norman 1994).

11. The First Instrumental Value of Political Representation • Objects or practices have instrumental value if they help achieve relevant extrinsic ends, i.e. objectives that are important independently of the objects or practices that helped achieve them (Réaume 2000). • The argument about the first instrumental value of political representation is based on the recognition that political institutions have an impact on the distribution of powers and resources among political actors (March and Olsen 1989). Those ethnic groups that are represented in formal political institutions can affect the distribution of resources and powers in their own favour (Pande 2003).

12. The Second Instrumental Value of Political Representation • The second instrumental value of political representation is based on the implications of the ethnic security dilemma as described by Saideman (1998) and Saideman et al. (2002): political representation makes ethnic groups feel politically, physically and economically more secure, as it increases the likelihood with which they can influence potentially harmful government policies.

13. A Grievance-Based Explanation (summary) • In sum, political representation is a valuable good for ethnic groups on three important accounts: (1) for its affirmation as recognised member of a political community, (2) the likelihood with which resources and powers are distributed in its favour, and (3) its perceptions of political, physical and economic security. • Grievances are therefore expected to rise among those ethnic groups who perceive that the design of political institutions reduces their chances of being represented politically. • As these grievances might translate into violent action, the odds of ethnic violence can be expected to increase under political institutions that provide only low chances of political representation. (See Theuerkauf 2010)

14. Part III: The Impact of Executive Structures on Ethnopolitical (In)Stability

15. Defining Presidential, Parliamentary and Mixed Forms of Government • The precise definition of parliamentary, presidential or mixed forms of government is contested (e.g. Sartori 1997). • Following the relatively unambiguous definition by Cheibub (2007): • systems in which the government cannot be removed by the legislature are presidential; • systems in which the government can only be removed by the legislature are parliamentary; • systems in which either the legislature or the independently elected president can remove the government are mixed.

16. The Virtues of Presidentialism in Divided Societies • An independently elected presidency may enable ethnic groups to gain access to executive power, even if they do not hold a majority in parliament (Horowitz 1991), and hence reduces ‘the stakes of control for any particular institution or office’ (Sisk 1996: 54). • Combined with a broad-based electoral system for the presidency, presidential systems encourage pre-electoral inter-ethnic coalitions (ibid.). • The separation of powers between president and legislature allows each to serve as a check on the other, so that ethnic groups have ‘more points within the system to block unfavorable actions’ (Saideman et al. 2002: 111).

17. The Virtues of Presidentialism in Divided Societies (2) • A president, elected by the whole nation as one constituency, and in his/her constitutional status as chief executive can serve as a symbol of national unity (Suberu and Diamond 2002). • A strong president (arguably) can overcome potential political deadlocks in divided societies (Horowitz 1985; Sartori 1997). • Presidential systems increase the identifiability, accountability and legitimacy of the political system (Shugart and Carey 1992).

18. Do These Virtues Really Hold? Not necessarily: • The president’s function as symbol of national unity relies on her statesmanship (Sisk 1996). • Presidential systems are not necessarily more effective forms of government (Sartori 1997) and are in fact politically more unstable than parliamentary systems (Cheibub 2007). • Not all presidential systems are based on a clear separation of powers (Alonso and Ruiz 2005). • If there is no broad-based electoral formula for the presidency, there is no guarantee for pre-electoral inter-ethnic coalitions; and even if these coalitions exist, they may be very short-lived (Suberu and Diamond 2002).

19. The Perils of Presidentialism in Divided Societies • Whereas parliamentary systems possess collegial cabinets (Lijphart 2004) and make coalition-building more likely (Cheibub 2007), presidential systems rely on a one-(wo)man-executive, and make political gains and losses absolute. • As only one person can become president, the zero-sum-game between presidential candidates is inevitable (Linz 1990). This can lead to polarisation in the populace as well as among politicians, exacerbated by the fact that the president’s office is temporally rigid so that political winners and losers are defined for her entire mandate (ibid.).

20. A Grievance-Based Explanation Applied to the Design of Executive Structures • In this manner (i.e. through the zero-sum character of presidential elections, the non-collegial nature of the executive, the lack of incentives for coalition-building and the temporal rigidity of the presidential office), presidential systems reduce the chances of political representation particularly for ethnic minorities. • According to the grievance-based explanation of ethnic violence outlined above, presidential forms of government would thus be expected to increase the odds of ethnic violence.

21. Part IV: Choosing Electoral Systems for Divided Societies

22. Majoritarian Electoral Systems • Majoritarian electoral systems, such as the single-member district plurality or majority system, are based on winner-take-all principles: The candidate with the largest share of votes is elected, while all other voters remain unrepresented (Lijphart 1999). • In legislative elections, this leads to an exaggeration of votes for the party that gains a nationwide majority or plurality of votes, with the aim to produce a decisive and effective parliamentary majority (Norris 2002). • Moreover, as DeVotta (2005) argues, the first-past-the-post electoral system in Sri Lanka also encouraged strategies of ethnic outbidding.

23. Proportional Electoral Systems • Electoral systems of proportional representation stand in contrast to the winner-take-all logic of majoritarian systems: by aiming to ensure the proportional translation of a party’s share of votes into the number of its seats in parliament, they lower the hurdles for smaller parties, increase the effective number of parliamentary parties and thereby heighten the number of possible political winners from legislative elections (Cohen 1997; Norris 2004). • Moreover, list PR is argued to encourage ethnically diverse party lists and to make inter-ethnic coalitions more likely by virtue of a more fragmented party system (Davis 2004).

24. Majoritarian vs. Proportional Electoral Systems • According to the grievance-based explanation of ethnic violence outlined, majoritarian electoral systems for the legislature should thus increase the odds of ethnic violence, as they reduce the chances for ethnic minorities to be represented politically. But it is not that simple… • See last week’s debate: some proportional arrangements have been criticised for cementing ethnic cleavages in the political system (esp. consociational arrangements). • Party lists need not be ethnically diverse in divided societies (Davis 2004). • There is no guarantee that inter-ethnic coalitions will be formed (ibid.).

25. Majoritarian vs. Proportional Electoral Systems (2) • Moreover, a highly fragmented party system could lead to political deadlocks. • List PR is not necessarily associated with higher levels of political support among ethnic minorities (Norris 2002). • Majoritarian electoral systems can still make special provisions such as seat reservations in order to accommodate minorities (ibid.).

26. What About Vote-Pooling Electoral Systems? • To make it even more difficult, there is also the debate about vote-pooling (preferential) arrangements. • Vote-pooling arrangements can either be majoritarian in nature (such as Alternative Vote) or proportional (such as Single Transferable Vote).

27. What About Vote-Pooling Electoral Systems? (2) • Vote-pooling electoral systems promote inter-ethnic pre-electoral coalitions, as they enable voters to rank candidates in order of their preference and hence encourage politicians to campaign not just for first-preference votes from their own community, but also for second-preference votes from other groups (Horowitz 1991; Reilly 2001). • In this manner, electoral systems such as AV or STV should provide incentives for politicians to ‘pool votes’ across ethnic lines, moderate their views on ethnopolitically controversial issues, and to seek inter-ethnic compromises (ibid.).

28. What About Vote-Pooling Electoral Systems? (3) • Vote-pooling (aka preferential) electoral systems thus foster centripetalism and avoid the danger of cementing ethnic cleavages inherent in list PR (Reilly 2001). • However, vote-pooling electoral systems frequently have been criticised for arguably being too complicated. • More importantly, there is simply not enough empirical evidence to properly assess the effects of vote-pooling electoral systems (Sisk 1996).

29. Part V: Summary

30. Summary • The debate about which form of government and electoral system for the legislature are most suitable for ethnically divided societies has remained inconclusive to date. • Whereas the separation of powers between the executive and legislature in presidential systems arguably has conflict-reducing effects, the inevitable zero-sum character of presidential elections, non-collegial executive structure and temporal rigidity of the presidential office reduce chances of political representation for ethnic minorities. • The main goal of majoritarian electoral systems is to produce decisive and effective parliamentary majorities. Proportional electoral systems, on the other hand, aim to increase the inclusivity of the political system. Vote-pooling electoral systems arguably avoid the danger of cementing ethnic cleavages.

31. Summary (2) • Following the grievance-based explanation of ethnic violence that identifies three crucial values of political representation, political institutions that reduce the chances of political representation for ethnic minorities should increase the odds of ethnic violence. • A parliamentary form of government and proportional electoral system (whether it relies on vote pooling or not) would thus seem to be most suitable for ethnically divided societies, as they prima facie increase the inclusivity of the political system. • Yet, each institution has its strengths and weaknesses, and there is no clear empirical evidence to favor one type of executive structure or electoral system for the legislature over the other. There is still a long way to go in the research on institutional incentives for ethnic violence.