Accommodating, Opposing, or Dismissing? Ethno-Regional Mobilization, (De)Centralization, and State-Wide Party Strategies in Nigeria and Kenya

Abstract Political parties in much of sub-Saharan Africa are often state-wide, but the societies are characterized by salient ethno-regional cleavages. State-wide parties, therefore, frequently encounter strategic challenges around ethno-regional mobilization and demands for special rights or self-rule. What shapes these parties’ strategic choices? We analyze whether party strategies are characterized by accommodation, opposition, or dismissal, and argue that their strategies are shaped by the degree to which contestation around territorial politics are centralized. We examine this argument by comparing how regionally dominant state-wide parties have responded to ethno-regional mobilization in the Niger Delta in federal Nigeria and the Coast region in more centralized Kenya over the last decade, and find that centralized contestation over ethno-regional issues forces parties to engage with ethno-regional demands while decentralized contestation allows parties to dismiss them.


Introduction
Pronounced ethno-regional cleavages are common in sub-Saharan Africa, and ethnic and regional identities are frequently mobilized around elections. While the relations between ethno-regional grievances and party politics are significant, they are far from straightforward. Recent research has demonstrated that, contrary to commonly held views, political parties in sub-Saharan Africa are often not mono-ethnic or purely regional. 1 Often parties are state-wide, participating in elections and having representation in a majority of regions in the national (or in the case of Nigeria: the federal) state. 2 At the societal level, however, most countries on the continent are characterized by salient territorial cleavages furthered by regional inequalities and sub-national territorial concentrations of identity groups. 3 This frequently translates into strong ethno-regional mobilization, including distinct grievances and demands for self-rule or preferential treatment. 4 Such demands, whether articulated by regional parties, social movements, or both, present state-wide political parties with the serious challenge of how to anchor or sustain support in various regions. This challenge revolves around balancing support for and from different constituencies when parties are faced with hard political choices on polarizing issues in competitive contexts. What shapes the strategies of state-wide parties facing ethno-regional mobilization in sub-Saharan Africa is surprisingly understudied. This article addresses the topic by analyzing how centralized ethno-regional contestation shapes party strategies.
While parties' specific strategic choices vary with the competitive contexts in which they are made, 5 research suggests that strategies are patterned by both structural and institutional features. The literature underscores the significance of territorial cleavageswhich holds for most societies in sub-Saharan Africa-and the structure of the polity. The latter dimension is mainly discussed in terms of the party system and the centralization of state structures, which have significant but complex relations to party strategies. 6 Much of this debate has been based on European cases and centered on the formal features of state structures and party systems. In many parts of the world, however, formal features may not be reliable indicators of where power and contestation are located. 7 This article therefore investigates how the actual centralization of contestation and authority shapes state-wide parties' strategies around ethno-regional mobilization in societies with highly salient ethno-regional cleavages. Previous research suggests that such strategies can be analytically categorized as accommodating, opposing, or dismissing ethno-regional mobilization depending on the electoral challenges facing the parties. 8 Drawing on this distinction, we explore the argument that party strategies on territorial issues are shaped by the level on which ethno-regional political contestation is played out-that is, the actual (de)centralization of that contestation. We further argue that the level of contestation in turn corresponds with the level of actual authority over the relevant issues. Where authority is centralized, contestation around ethnoregional issues by social mobilization and party-political competition is likely also to be centralized. If ethno-political cleavages are articulated at the level of national politics, pressure on state-wide parties to engage with (either accommodate or oppose) ethnoregional mobilization will be strong, and the cost of ignoring them higher. If, however, such politicization is articulated mainly at the sub-national political level, it is easier for state-wide parties to ignore pressing demands.
The article explores this argument through a comparative analysis of the strategies of regionally dominant state-wide parties in relation to ethno-regional mobilization in two sub-national regions with pronounced demands for strong forms of self-rule: the Niger Delta region in Nigeria and the Coast region in Kenya. In both countries, ethnoregional cleavages are historically rooted, deep-running, and salient, and the territorial dimension is significant in political competition. The countries differ, however, in terms of how cleavages and competition have been institutionalized through territorial state structures, with implications for the level of centralization of contestation around territorial politics. Kenya has long been heavily centralized and has only recently embarked upon a policy of moderate devolution, while Nigeria is a federation built on ethnoregional parts; federalism was introduced with the aim of diverting ethnic conflict from the level of national politics. The difference between the countries in terms of the degree of centralization is not merely formal: though Nigeria is not as decentralized in practice as its formal features indicate, it is still less centralized than Kenya, particularly in terms of ethno-regional mobilization. The article makes several contributions to existing research. Theoretically, it links party strategies on the territorial dimension of competitive politics to the level of centralization of authority and contestation.
Empirically, it brings cases from sub-Saharan Africa, hitherto understudied in this respect, to the literature on party strategies. The next section reviews the literature on state-wide parties' strategies around territorial politics, develops the argument about how the level on which contestation and authority are actually played out shapes party strategies, and outlines the comparative design. The argument is examined empirically in the subsequent two sections, where we analyze how the dominant state-wide political party in each region-the People's Democratic Party (PDP) in the Niger Delta and the Orange Democratic Movement (ODM) on the Kenyan Coast-has engaged with ethno-regional movements and demands. We situate the cases in historical context, but focus the analysis on the last decade, with an emphasis on two recent election cycles in each country. The empirical analysis, which draws on media material and interviews with independent analysts and representatives of political parties and ethno-regional movements, centers on how parties strategize to face demands for self-rule and preferential treatment. The final section concludes by assessing the strength of the argument.

Territorial politics and state-wide parties
Earlier literature on relations between political parties and ethno-regional groups and movements in sub-Saharan Africa has focused on the role of distinctly ethnic parties and ethnic outbidding, by which politicians competing for the support of an ethnic group push one another in ever more extreme directions. 9 Recent research, however, has demonstrated that while ethno-regional cleavages are generally important in electoral politics in sub-Saharan Africa, their significance differs across countries and affects politics in indirect ways. 10 Parties in ethnically plural societies are not necessarily distinctly ethnic in the sense of having their support based only in one ethnic group. Because, among other reasons, demographic ethnic fragmentation necessitates coalitionbuilding, parties often are broader in scope and reach and sometimes are state-wide-in themselves, as parts of national coalitions, or as so-called ethnic congress parties of ethnic elite alliances. 11 While this topic is understudied in sub-Saharan Africa, more has been written on other geographical regions, and there is an extensive literature on how state-wide parties in Europe engage with sub-national challenges. 12 Much of this literature acknowledges the continued relevance of the sociological perspective of Seymour Martin Lipset and Stein Rokkan 13 on how cleavages shape party systems, 14 but also stresses factors such as territorial state structures in shaping party competition. For example, decentralization compels political parties to adapt in order to deal with new arenas of competition and the emerging issues, resources, and actors these arenas create. 15 While relations between state structures and party politics are deemed significant, the precise mechanisms are contested. Pradeep Chhibber and Ken Kollman argue that authority migration shapes party systems: over time, centralization of state structures promote nationalization of party systems and party-political competition. 16 Others point to examples of reverse causation, as when political parties, following deepened ethno-regional cleavages, promote decentralization. 17 As a first step in addressing the research question, we build on the general proposition that the level at which authority is exercised shapes party politics. However, we specify existing arguments in order to capture the scale of actual authority. First, while we acknowledge that the locus of authority is linked to formal territorial state structures, we do not assume it is determined by it. Sometimes, decentralization is merely nominal and does not involve substantive devolution of power and resources. Second, the nationalization of the party system may not indicate the level of contestation. For instance, the prevalence of ethno-regional parties in Kenya does not mean that power is decentralized, but that territorial cleavages have been more salient than state structures and the electoral system in shaping the party system. Ethno-regional grievances in Kenya tend to be channeled toward and articulated at the level of national politics by ethno-regional movements and political parties.
Theorizing party strategies and the scalar level of contestation Our argument proceeds in two steps. We first build on existing theories on strategic party behavior and then relate these to the level of contestation. To analyze party strategies we draw on the typology developed by Bonnie Meguid, who argues that mainstream parties, when faced with electoral threats from niche parties, choose between dismissive, accommodative, and adversarial strategies. 18 Her position, salience, and ownership (PSO) theory of party behavior posits that parties develop strategies by positioning themselves on issues, seeking to change the salience of existing and new issues, and competing for ownership of them. We apply the core features of Meguid's argument to examine state-wide parties' strategies toward ethno-regional mobilization in societies with high-salience ethno-regional cleavages. We understand ethno-regional mobilization in a broad sense, as articulated by regional parties, social movements, or both. Even in the absence of an ethno-regional party, an ethno-regional movement may challenge the role of a regionally dominant state-wide party as the main vehicle for channeling regional interests. A further challenge for state-wide parties arise when, as in the cases under study, the regions are ethnically heterogeneous, and where it can be assumed that opinions about ethno-regional mobilization differ among constituencies, including moderates and radicals within both majority ethnic groups and minority communities.
A first distinction is between strategies that engage with competitors and those that dismiss them. 19 If ethno-regional mobilization poses only a modest challenge, the statewide party is likely to dismiss it by keeping silent or downplaying its importance, to reduce the salience of ethno-regional issues, or to retain it, but at a low level. When such a challenge is stronger, state-wide parties need to engage with it. When the statewide party is ideologically close to the ethno-regional actor or demand, it will seek accommodation by close positioning to claim ownership of the issue. When it is ideologically distant from the challenge, it will seek an adversarial strategy and distinctly position itself far from the issue. Both these strategies will raise the salience of ethnoregional issues. 20 Two cases that illustrate the argument are Spain and the United Kingdom, where the growth of Catalan and Scottish nationalist movements from the 1970s onwards constituted electoral threats to the Spanish Socialist Party (PSOE) and the British Labor Party respectively. The parties responded to these challenges by shifting their positions on territorial politics in favor of devolution, whereas the conservative parties in both countries remained opposed to such accommodation. Later, however, all state-wide parties in both countries opposed calls for independence. 21 In our analysis, we make use of this categorization of strategies; for the sake of linguistic convenience, we label them accommodating, opposing, and dismissing.
Turning to the second part of our argument, we assert that state-wide parties' choices of strategy are conditioned by the extent to which political contestation around ethnoregional mobilization is centralized. Here we refer to the level on which contestation is played out, which can range from local to national. While ethno-regional mobilization is always locally anchored and sometimes has local origins, it may nevertheless primarily target and be articulated at the national level. When this is the case, and when responses refer to national politics, we regard contestation as centralized. We argue that the level of contestation corresponds with the level of political authority. In societies characterized by centralized authority and ethno-regional cleavages of high national significance, ethno-regional movements and political parties are likely to channel grievances directly to the national level so that both ethno-regional mobilization and the party political competition around it are centralized; in societies with decentralized authority the reverse can be expected. The main reason for this is straightforward: the locus of resources and decision-making power attracts contestation. Another reason is that in decentralized systems, the coordination of mobilization is more complicated, and more grids prevent it from happening. 22 We also expect party strategies to be shaped by the level at which contestation occurs. If ethno-regional contestation is centralized, its mobilization will constitute a significant challenge to state-wide parties that have or seek an electoral base in that region, and the pressure to engage with the challenge will be higher. The strategy of the state-wide party will be to accommodate or oppose it, depending on ideological proximity. 23 If the ethno-regional contestation is decentralized, the challenge to the state-wide party will be reduced and the most likely strategy will be to dismiss it. This is parallel to findings on ethnic conflict and decentralization that suggest that substantive decentralization tends to disperse conflict from the center to sub-national arenas. 24 Finally, we consider the two main alternative explanations for party strategies, both in the analysis of each case and in our conclusions. One alternative is that parties in power are more likely to dismiss or oppose demands that seem to undercut the influence of the national government, while opposition parties may recognize such demands as a way to access power. The other alternative explanation is that more coherent and efficient movements will have a greater leverage to force parties to accommodate their demands, while fragmented and feeble movements will have limited prospects of extracting concessions.

Comparing regions and parties in Nigeria and Kenya
To explore our argument, we compare the strategies of regionally dominant state-wide parties toward ethno-regional mobilization in two regions: one in federal Nigeria and one in centralized Kenya. The study covers two general election cycles in each country. 25 There are important similarities between the countries in the general political significance of territorial politics. In both countries, regional inequalities and subnational identity politics revolving around region and ethnicity translate into ethnoregional grievances and mobilization around marginalization. Electoral processes are intensely contested and closely fought, and political parties and coalitions need to balance support for and from different constituencies. Both countries practice a majoritarian electoral system and have laws against ethnic political parties. The two regions also exhibit similarities. Both are sites of deeply rooted grievances around economic marginalization and political exclusion recently expressed by social movements. At the same time, both regions lack strong specific ethno-regional political parties and are of comparable importance in national elections. The Niger Delta is where much of the oil that is central to Nigeria's economy is found, and the Coast region is a major tourism hub and home of Mombasa Harbor.
Cleavages and competition, however, have been institutionalized differently through territorial state structures. The implications of these institutional differences on the centralization of authority and contestation around ethno-regional issues allow a comparison of this dimension. In Nigeria, federalism has decentralized ethno-regional political mobilization, while in Kenya, the political system has created a legacy of centralization. Although the Kenyan constitution of 2010 introduced devolution to 47 counties, the allocation of power and resources to sub-national units is still modest compared with the budgets and decision-making authority vested in the 36 Nigerian states. Nigeria's revenue distribution formula requires 47.3% of national revenues to be transferred to state governments (with 20.6% to be forwarded to local government councils). Combined with a derivation principle giving 13% of natural resource revenues to the state of origin, some of the oil-rich states in the Niger Delta have budgets larger than those of many small African countries. The federal units have the power to make laws and establish courts, but the state governors have executive powers. Kenyan devolution is constitutionally anchored and thus on a firm footing, but other aspects point in a different direction. Only 15% of national resources are transferred to the counties, and the political authority of the counties is limited to less important issues. 26 Based on our stated argument, we therefore expect that decentralization of political power and resources in federal Nigeria makes it likely that dominant state-wide parties there will dismiss ethno-regional mobilization, while in Kenya, state-wide parties are likely to engage with (accommodate or oppose) such mobilization due to the a more significant challenge they pose on the national level.
The study is based on extensive field research. Fieldwork for the Kenyan part was conducted in the Coast region in 2018 and centered on 36 semi-structured interviews with leaders and members of political parties and ethno-regional social movements, community leaders, civil society activists, and local political analysts. The interviews took place in venues of the respondents' choice. They typically lasted for around one hour and revolved around the characteristics of ethno-regional mobilization and political party responses. In most cases they were conducted in English, but an interpreter was always at hand to ensure that conversations could be held in Kiswahili whenever respondents so wished. In Nigeria, fieldwork was conducted at various times from 2002 to 2018 and included 20 interviews with politicians, social movement activists, and community leaders. Issues raised in the semi-structured interviews are related to political and electoral mobilization, the links between political parties and society, and relations between ethnic groups. The interviews typically lasted between 45 and 90 minutes. In both countries, interviewees were selected using a snowballing technique with multiple entry points. Along with the interviews, the study builds on written documentation. We conducted a systematic search in Factiva, a global news database covering local media and international news wires, for reports beginning in 2005 retrieved using relevant keywords. 27 In addition to media material, to capture party strategies we examined (when available) party manifestoes, election campaign materials, and to identify the main dimensions of ethno-regional mobilization, we studied speeches and statements by various groups who mobilize around ethno-regional issues. We also drew on a rich secondary literature. When analyzing the material, we concentrated on the most significant and sustained ways in which political parties related to key grievances and demands voiced by ethno-regional movements to distinguish salient demands and responses from more temporary and isolated pronouncements. Similarly, when classifying party strategies in terms of accommodation, opposition, or dismissal we identified the more systematic and long-standing features of party pronouncements and actions. Both direct and indirect questions were asked during the interviews to corroborate and develop the classification of party strategies. In the Kenyan case, all respondents agreed on the nature of ethno-regional grievances and demands over time, and most had a similar assessment of the character of party strategies, though the explanations differed slightly. In the Nigerian case, respondents were suspicious of the parties' intentions and typically dismissed parties as platforms for personal ambitions and patronage disbursement structures. Occasionally, parties pursued one strategy toward the radical movements themselves and another toward the issues they voiced; when this was the case, we discuss the relative significance of each strategy.

Demands for territorial autonomy in the Niger Delta, Nigeria
The Niger Delta region is in the south-south geopolitical zone of Nigeria and includes the states of Bayelsa, Delta, and Rivers. 28 The main ethnic group in the area is the Ijaw, but the region is also populated by many different ethnicities. The grievances of the Delta region have long revolved around violations of social, environmental, and political rights in the wake of oil extraction: the region is the primary site for Nigeria's oil extraction, accounting for 70% of national revenues. 29 Popular demands call for the transfer of ownership of natural resources to the communities where they are located or "resource control." 30 The Movement for the Survival of the Ogoni People (MOSOP), the Niger Delta's "prototypical social movement," 31 has consistently demanded the federation be composed of self-governing ethnic states so as to ensure "political autonomy to participate in the affairs of the Republic as a distinct and separate unit." 32 While the political parties have largely dismissed Niger Delta grievances, they have also sought to capitalize on the language of resource control, but stress instead the need to transfer oil revenues to the states. 33 Ever since the inaugural elections after military rule in 1999, the Niger Delta has been a PDP stronghold. 34 The party, which won the 1999 elections, has not based its dominance in the area on accommodating ethno-regional demands. While introducing its manifesto with a statement on the need for "true federalism," PDP has not had a substantive agenda for relocating territorial powers. 35 PDP governors in the Niger Delta states support the national status quo of federal ownership of the natural resources and "full control of all the resources of Niger Delta." 36 Several institutional features of the Nigerian polity contribute in keeping ethnoregional demands away from the national center and in promoting state-wide parties. Nigerian federalism is marked by its stated aim to mitigate national conflict between different ethno-regional interests. 37 Political party regulations prohibit parties that lack cross-regional representation. 38 To win a presidential election, the successful candidate must attain at least 25% of the votes in two-thirds of the states. An informal principle of "zoning and rotating," initiated by PDP and subscribed to by all major parties, ensures that political and administrative office holders come from different parts of the country or the state. The principle operates as a power-sharing mechanism between different groups and is intended to reduce ethno-regional stakes in politics.
These mechanisms deflect ethno-regional demands from national politics and refer them to the sub-national level. Another factor that adds to the decentralization of ethno-regional demands is complaints by oil-producing communities to state governments about lack of consultation or inclusion around allocating projects and determining budgets. 39 Sub-national political contestation often revolves around which ethnic group has the right to claim political positions and appointments within a designated territory. Such conflicts have been fueled by state intervention through both gerrymandering electoral wards and distributing oil rents in the form of development projects. 40 Allocating benefits, dismissing demands: the 2011 and 2015 elections in the Niger Delta region Political events beginning in 2007 created new opportunities for PDP to relate to Niger Delta grievances. The 2007 elections produced an Ijaw vice-president from Bayelsa State, Goodluck Jonathan. The PDP presidential candidate, Yar'Adua, argued that a vice-presidential candidate from the Niger Delta was a way of "building confidence" in the region in light of widespread violent and nonviolent protests against oil exploitation. 41 It is, however, more likely that the choice of a running mate from the Niger Delta can be attributed to the zoning and rotating principle, according to which the vice-presidency had to go to the south-south geopolitical zone. The choice of Jonathan as vice-presidential candidate also enabled PDP to strengthen its grip on the Niger Delta without accommodating the demands for resource control or territorial autonomy. Further benefits, however, were allocated to the region. In 2008, the government created a Ministry of Niger Delta Affairs, and in 2009, militants who surrendered their weapons and renounced militancy were offered an amnesty including an unconditional pardon and stipends. However, neither the amnesty program nor having a vicepresident from the region addressed the demands for increased territorial autonomy. 42 The new institutions and socio-economic programs can be seen as continuations of the strategy to co-opt the militant protesters while dismissing demands that challenge power structures and resource allocation. 43 Ahead of the 2011 elections, PDP entered the campaign with Jonathan as incumbent president following the death of Yar'Adua the year before. The major opposition party, Congress for Progressive Change, drew its main support from the north and had limited presence in the Niger Delta, whereas PDP had the presidential candidate and controlled the gubernatorial offices. Regional demands for devolution and territorial self-determination were ignored. PDP incorporated regional politicians and opinion leaders without making any promises to seek increased territorial control for ethnic communities. The Niger Delta states were effective in swelling the vote. Jonathan was reported to have attained over 97% of the votes in Cross River, Rivers, and Delta states and 99.6% in his home state, Bayelsa. Overall, the result was an indication more of the level of control that PDP retained over state institutions in the region than of any popular support for its political program.
Up until 2013, PDP was the only real state-wide party in Nigeria. While opposition parties had a nominal national presence, they had limited support outside their strongholds, making it easier for PDP to dismiss Niger Delta demands. The opposition, however, changed its strategy ahead of the 2015 elections as the three major opposition parties merged to form the All Progressives Congress (APC). Rather than challenging PDP in the president's home region, APC chose to concentrate its efforts in the north and the south-west. Nevertheless, one state in the Niger Delta was brought into the national limelight. Jonathan's candidacy for another term had been controversial for many within PDP, and several governors and members of the National Assembly defected to APC before the elections. Among the defecting governors were Rivers State's Rotimi Amaechi. This provided an opportunity for APC to get a foothold in the Niger Delta, and Amaechi gained a central role in APC as presidential campaign director. 44 This did not, however, mean that Niger Delta demands were made part of the election campaign. To the limited extent that regional issues were addressed in campaigns and party manifestoes, the mentions were general and non-committal in terms of environmental degradation, unemployment, and poverty. 45 Both APC and PDP were silent on ethno-regional demands.
The scope for developing a common agenda in the Niger Delta is weakened by a movement that is fragmented in both identity and strategies; while some groups adhere to nonviolent methods, others make use of violence. This fragmentation has been exploited not only by the national parties, but also by local politicians, who more often than not speak the language of resistance, but in a way that makes it compatible with existing structures. 46 Governors and state legislators in the Niger Delta get their positions by relying on support from central party structures. In that process, political aspirants draw on the vocabulary of ethno-regional social movements but dismiss their substantive demands. Can such strategies be explained by whether the party is in government or in opposition? It does not seem to be the case; both parties largely pursued dismissal strategies. What about the strength of the movements? It is evident that a more cohesive protest movement would have created a greater challenge to political parties, but it is not possible to say whether this would have created a change in strategy.

Ethno-regional grievances and political parties in Kenya's Coast region
Kenya's Coast region is relatively underdeveloped economically and politically marginal. Since independence its main articulated regional grievances have been landlessness, unemployment, and political marginalization. 47 Kenya's eight provinces, of which Coast Province was one, were abandoned as administrative entities following the establishment of counties at the 2013 elections, but popular parlance still refers to them. In political terms, the core of the Coast is the four littoral counties Mombasa, Kilifi, Kwale, and Lamu, which share a history slightly different from the other two Coast counties, Taita Taveta and Tana River. With colonialism, a ten-mile strip inland from the Indian Ocean was carved out from the British Empire and made a part of the Sultanate of Zanzibar. At independence, the different communities in the area (Arab, Swahili, and various African ethnicities coming together as Mijikenda) were divided over the region's political future, though no group was willing to join a centralized Kenyan state. The Arab and Swahili communities in the Coast People's Party demanded independence for the Coastal strip, while the Mijikenda organized in the Coast African People's Union, soon incorporated into the Kenya African Democratic Union (KADU), which pushed for the integration of the Coastal strip into a federal Kenya. 48 The KADU line held sway, but the 1963 federal independence constitution was abandoned the following year. As a consequence of this and of the incorporation of KADU into the governing Kenya African National Union (KANU), the political voices of the Coast became weaker and more fragmented; their deepening social grievances about land, unemployment, and lack of influence could find no forceful or united political outlet. 49 The country's first two Presidents, Jomo Kenyatta (1963-78) and Daniel arap Moi (1978Moi ( -2002 skillfully cultivated sections of the region's political leadership to contain the discontent and to redirect ethno-regional grievances onto local minorities associated with the opposition instead of making demands on the national government. 50 The reintroduction of multi-party politics in 1991 made the Kenyan political landscape more fluid and sub-national regions became both topics and arenas of intensified politicization. KANU succeeded in portraying itself as the protector of smaller communities against the threat of upcountry ethnic groups, which at the time was associated with the opposition. In both the 1992 and 1997 elections the majority Coastal vote was for KANU and Moi, although the opposition performed better in 1992 than in 1997. 51 Political tides turned in 2002 when the Coast became part of the National Rainbow Coalition that voted KANU out of power, and in the negotiations over a new constitution during the first few years under the new government, Coastal delegates made farreaching demands about resource allocation and regionalism. 52 In the 2005 referendum, the Coast region solidly rejected a proposed new constitution that, contrary to promises made in 2002, would entrench centralized rule. From the referendum campaigns onwards, national politics were strongly shaped by divisions over grievances around marginalization and exclusion. 53 To a large extent, these centered on the issue of devolution. In the 2007 elections, the Coast region by and large joined forces behind the opposition candidate Raila Odinga and his ODM. Since its formation in 2006 ODM, like many other parties in Kenya, has had its most loyal base in one region: in this case Nyanza, the home region of Odinga's ethnic group, the Luo. Unlike most other Kenyan candidates and parties, however, over the last three elections both Odinga and ODM have gained the vast majority of their votes and elected representatives from other parts of the country. Throughout, ODM has had a major presence in all parts of the country except for Central Kenya and the Rift Valley (after 2013). The party has nation-wide local branches with elected leaders and a national leadership drawn from many parts of the country. In 2007, ODM campaigned on introducing devolution and addressing land rights and historical injustices. These were issues that spoke directly to (and even about) Coastal grievances by addressing the role of the Coast Province within Kenya as well as the plights of Coastal communities within Coast Province in terms of land, employment, and economic opportunities. 54 To summarize, national politics from 2005 onwards changed regional politics on the Coast. It raised the stakes, brought Coastal issues into national politics, and heightened expectations.
Accommodating devolution, opposing secessionism: Coastal radicalism and ODM strategies for the 2013 elections After the 2005 referendum, the Coast region leaned heavily towards ODM, mainly due to the party's stance on devolution. The inclusion of devolved state structures in the 2010 constitution thus ought to have strengthened ODM. However, from 2010 until the 2013 elections, the political ground in the Coast region shifted. Coastal grievances were given unprecedented attention and cohesion, at least discursively, through the Mombasa Republican Council (MRC), a social movement whose main demand was the Coast region's secession from Kenya. 55 The MRC was formed in 2005 in response to the upcoming referendum on the constitution and debates about devolution, 56 and it mobilized around the perennial problems of landlessness, resource distribution, and political marginalization. The rapidly growing attention to the MRC from around 2010, when it began to gain traction, suggested that it could galvanize Coastal communities and channel long-standing grievances into energized demands, advanced mainly in the form of court cases. 57 The rising popularity of the MRC coincided with the criminalization of the organization in 2010 on the grounds that it carried out armed attacks, something the movement always refuted. The High Court lifted the ban on the MRC in 2012, but its leaders and members continued to be targeted by the police. The government crackdown was, in stages, both a cause and a consequence of the MRC's call since late 2011 for a Coastal boycott of the 2013 general election. 58 The MRC's secession demands and electoral boycott brought Coastal grievances to the level of national politics 59 and posed a serious threat to ODM's electoral aspirations, all of which forced the party to choose a strategy of engagement. 60 This was reinforced by the MRC's ability to capitalize on both the perceived weakness of regional members of parliament and the disappointment in some quarters that devolution according to the 2010 constitution would bring not regional self-rule, but merely a weaker form of service-delivery oriented decentralization. 61 The MRC regarded this as another sign of the inability of compromised local political elites to fight for regional priorities. 62 The ethnically fragmented nature of ODM's Coastal constituency made the party pursue a combined strategy of opposition and accommodation. 63 It marked its distance from the MRC's calls for secession, while at the same time refraining from condemning the movement and possibly losing the support of moderate MRC sympathizers. 64 Meanwhile, ODM began to appropriate much of the MRC's core message. 65 The party therefore campaigned on issues that would resonate with Coastal grievances, with a strong emphasis on devolution and land issues. 66 ODM contrasted the MRC's exit option with a more realistic chance of ODM winning the presidential election. For ODM, the elections at the Coast brought mixed fortunes. While the MRC's call for a boycott did not have a full-scale effect, voter registration and voter turnout were lower on the Coast than in other regions. 67 Nevertheless, the Coalition for Reforms and Democracy (CORD) alliance, and ODM in particular, carried the day among those who did vote by tapping into general Coastal grievances and directing them toward capturing state power rather than refuting it through secession.
A State-Wide party takeover of regional grievances: the 2017 elections at the Coast Following the 2013 elections, Coastal grievances remained, but their content and the conditions for their political articulation had changed. The controversial 2013 presidential elections, regarded by the opposition as manipulated, left Kenya deeply divided and the opposition-leaning parts of the country feeling wronged, adding another layer to Coastal narratives of exclusion. Responses to marginalization, however, were different to those ahead of the 2013 elections. Following the severe government crackdown, the MRC was much weakened, and its leadership had either disappeared or was embroiled in court cases. 68 Furthermore, the introduction of devolved governance structures by way of counties created a new context for politicization. On the Coast, devolution reshaped the political articulation of regional issues in contradictory ways. While regional mobilization was complicated by new intra-regional political arenas with distinct issues and actors, local politicians could use the new platforms and resources to build their status in relation to national politics by confronting the national government and repositioning themselves within the opposition. As Ngala Chome argues, devolution did not conclude the debate on the marginalization of the Coast region within Kenya, it intensified it. 69 By and large, political authority in Kenya remained centralized, as did contestation around ethno-regional demands.
In this new context, how did ODM relate to Coastal grievances? It needed to cultivate continued electoral support by accommodating regionally anchored demands. After the 2013 elections, Coastal demands were increasingly seen to be promoted by the emergence on the national political stage of Ali Hassan Joho, Mombasa governor, deputy party leader of ODM, and a leader in the new opposition coalition the National Super Alliance (NASA). ODM in general and Joho in particular had taken over the grievances articulated by the MRC ahead of the 2013 elections and addressed notions of marginalization and rejection with defiance. 70 Joho's fearless leadership style resonated with a mass constituency of opposition sympathizers, not only on the Coast, and for the first time in over a decade a Coastal politician had elevated himself to the status of national figure. 71 In the process, Coastal issues of devolution, employment, and-to a lesser extent-land had been further incorporated into the ODM platform. 72 The strong emotional content of the campaigns made substantive issues even more resonant, 73 facilitated a common regional ground in distinct opposition to the center, and provided a temporary answer to the question of what would constitute a Coastal political identity given the ethnic, racial, religious, and territorial fragmentation of the groups in the region.
The 2017 elections confirmed ODM's solid grip on most parts of the Coast. In the August elections (among which the presidential election was eventually annulled due to irregularities and illegalities), the opposition gained the majority of votes and seats in the region, 74 though for obvious reasons the official figures from those elections cannot be regarded as entirely reliable. In the repeat presidential election in October, the region by and large adhered to the opposition boycott, and turnout was minimal. 75 Thus, the region supported ODM for as long as the party was seen to represent Coastal concerns and for as long as the Jubilee government was deemed hostile to the same issues. 76 The elections deepened Coastal disillusionment with national politics, and in November two ODM governors-Ali Hassan Joho of Mombasa and Amason Kingi of Kilifi-suggested that the region ought to secede from Kenya. 77 While this statement was generally regarded as a bargaining position rather than a demand to be taken literally, it illustrated both the depth of Coastal grievances and the fact that ODM had taken ownership of them.
Can these party strategies be explained by whether the party is in government or opposition? This is not evident. In 2013 ODM was formally in a grand coalition government and in 2017 it was in opposition, but throughout it pursued a strategy of accommodation. Ahead of 2017, the governing Jubilee Party (JP) engaged the Coast as a region by opposing regional demands for extended devolution and instead framing reasons for the region to support the JP in terms of infrastructural development. The strength of the ethno-regional movement is also a minor contributing factor, not a decisive one. The MRC captured imaginations and shaped the discourse in a strong way, but it lacked the organizational strength to pursue its demands, especially in the wake of the state crackdown. The trajectory of ODM's accommodative strategy towards Coastal grievances demonstrate that they were given more urgency by the MRC, but not decided by it.

Conclusions
This article examined how the level on which territorial politics are contested shapes the strategies of state-wide parties toward ethno-regional mobilization. We argued that in political orders with centralized authority, ethno-regional mobilization and partybased contestation will be channeled toward and articulated at the national level, and state-wide parties will engage with such mobilization, while in decentralized systems, state-wide parties are more likely to dismiss it. We supported our argument with a comparative study of these dynamics in the Niger Delta region in Nigeria and Coast region in Kenya. In the Niger Delta, the regionally dominant party, the PDP, has largely sought to dismiss ethno-regional movements and demands. On the Kenyan Coast, the regionally dominant ODM party made inroads in the region by accommodating key regional grievances such as demands for strong versions of devolution, national redistribution of resources, and land reform.
The argument is supported by other empirical examples in the two countries. In south-eastern and south-western Nigeria, ethno-regional movements have agitated for increased autonomy. Just as in the Niger Delta, state-wide parties have diverted the demands to the sub-national level. In the former Western province of Kenya, ethno-regional grievances around marginalization and demands for a fairer access to state power and redistribution of resources have been significant over the last decade, and ODM has sought to incorporate the region into its electoral coalition. In doing so, it has faced competition from three ethno-regional opposition parties in parliamentary elections, and it has therefore been pushed to accommodate and emphasize ethnoregional demands in order to gain regional support. 78 In Nigeria, the experiences of the civil war in the late 1960s were fundamental in shaping a political system with the aim of avoiding violent conflict over central control of state power. A key aspect of this has been federalism, designed to steer conflicts away from the national level and relatively successful in doing so. Another is the current party system, dominated by state-wide parties. This is conditioned partly by constitutional clauses and party legislation to ensure the national character of political parties and partly by parties' internal informal arrangements of zoning and rotating, aiming to distribute power throughout the federation. This has turned Nigerian political parties into state-wide vehicles that accommodate various regional interests and has made exclusive ethno-regional mobilization more difficult in national politics. The geographically broad character of the political parties make it easier for them to dismiss ethno-regional demands voiced by social movements. Instead, regional demands are incorporated into parties by regional political elites for internal bargaining purposes. While this has not eliminated the significance of ethno-regional issues in Nigerian politics, as illustrated by the intense ethnic struggles in the Niger Delta, it has devolved much ethno-regional contestation to the sub-national level.
In Kenya, politics have historically been more centralized than in Nigeria. The combination of deep-seated inequalities between regions and ethnic groups and the makeup of state structures and the political system has promoted a zero-sum character of politics centered on winning presidential elections, with ethno-regional groups mobilized into coalitions. The centralized character of Kenyan politics has thus created and institutionalized strong and direct links between national politics and ethno-regional mobilization and given great importance to the latter. While the 2010 constitution introduced devolution to address the problems of centralized power and the corresponding exclusion of communities, the effects in terms of decentralized political contestation are so far modest. The experiences of the 2017 elections suggest that institutional reforms have not yet changed the fundamental conditions of Kenyan politics. Kenya continues to be shaped by centralized contestation based on deep ethnoregional cleavages. The Presidency remains by far the most significant asset, and to gain it political parties continue to promote and conduct mobilization of ethnic and regional grievances.
What are the alternative explanations for these party strategies? One would center on whether the party is in government or opposition. While not the focus of our investigation, this factor does not seem to have affected the outcomes in the cases under study. In the Nigerian case, both state-wide parties pursued strategies of dismissal, and in Kenya, both state-wide parties embarked upon strategies of engagement, one by accommodation, one by opposition. It is certainly possible that parties in power will be more likely to dismiss or oppose demands that seem to undercut the influence of the national government, such as calls for far-reaching self-rule, but this was not shown in our data.
Finally, the strength of a protest movement obviously matters for party responses, in general. However, our analysis shows that in the Kenyan case, the changing strength of the MRC over time affected ODM's strategic choices only by degree, not by kind. While the movements in the Niger Delta region are relatively fragmented, they occasionally display a degree of unity by creating coalitions around demands, but again, this does not seem to alter the parties' fundamental responses.
Theoretically, the article adds support to the strand of the literature on party strategies that stresses the importance of relating strategies to institutions and structures. It contributes by pointing to the need to identify the political levels of contestation and authority. In many parts of the world, discrepancies abound between nominal and actual decentralization of political authority and between the formal appearance and the daily operation of party systems. These contradictions make it difficult to apply analytical frameworks and comparisons that rely mainly on the formal features of those factors. 77. Mohamed Ahmed, "Hassan Joho, Amason Kingi Call for Breakaway of Coast," Daily Nation, 3 November 2017, https://www.nation.co.ke/news/Hassan-Joho-Amason-Kingi-call-forbreakaway-of-Coast-from-Kenya/1056-4172150-2rxmh2z/index.html (accessed 3 November 2017). 78. In 2017, this was facilitated by the fact that two of the three regional parties were, together with ODM, members of the NASA coalition and united around Raila Odinga as their presidential candidate.