Traditional Leaders and Ethnic Politics in
Mozambique
Jasper van Doeselaar 10677453 Bachelor Thesis Political Science University of Amsterdam Research Project Civil War and State-Building Supervisor: Dr. A. A. Steele Second Reader: Dr. U. Daxecker Words: 8362 Date 30-01-20172
Abstract
This thesis intends to explore if there is sub-national variation in the strength of pre-existing hierarchical social structures in Mozambique and whether this affects political affiliation in the country along ethnic lines. The theory is that there is an inverse relationship between the strength of hierarchical structures and the degree of ethnic politics in a country. For the analysis, quantitative and qualitative evidence was used to compare the southern, central, and northern regions of the country. The findings suggest that there are particularly strong differences between the south, and the central and northern regions. In Mozambique, it seems that weak traditional hierarchical structures do not lead to more ethnic politics rather than the opposite. Stronger traditional systems lead to increased political salience of ethnicity. However, there is not enough evidence to suggest that this translates into a preference for one political party over another.3
Tabel of Contents
Abstract ... 2 Tabel of Contents ... 3 Introduction ... 4 Literature Review: Ethnic Identity and Political Competition in Africa & How Intermediaries Fit In ... 5 Theory: The Role of Local Intermediaries in Mitigating Ethnic Politics ... 6 Qualitative and Quantitative Evidence ... 8 Historical Legacy ... 8 Colonial Period ... 8 Emergence of Frelimo ... 9 Postcolonial government ... 10 Since 1992 Peace Accords ... 12 Survey Data Comparison ... 14 Testing the Ethnic Mechanism ... 16 Intermediaries ... 18 Analysis & Discussion ... 19 Conclusion ... 20 References ... 22 Appendix I ... 23 Appendix II ... 24
4
Introduction
Research Question: How do pre-existing hierarchical social structures explain political affiliation within Mozambique along ethnic lines?
This thesis intends to explore if there is sub-national variation in the strength of local intermediaries in Mozambique and whether this affects party affiliation in the country along ethnic lines. Theories suggest that in countries where local intermediaries play an important role during elections – which refers particularly to local leaders such as traditional leaders that are kin-based (eg. chiefs) and religious leaders (eg. reverends or priests) – the likelihood that political candidates will appeal directly to groups based on ethnicity is reduced (Koter 2013). A large body of literature and research on politics in Africa has focused on the relationship between ethnic identity and support for political parties or leaders, the assumption being that there is often an interaction between the two.
The political landscape of Mozambique is not characterized by ethnic blocs, but very much defined by the competition between the country’s two largest parties: Frelimo and Renamo. Their supporters and members cut across ethnic lines. However, Frelimo is particularly popular in the south of the country and in the two most northern provinces whereas Renamo traditionally has the most support in the more central regions. This electoral pattern is a legacy of a civil war that ravaged the country for fifteen years from 1977 till 1992, shortly after independence in 1975. Both parties started out as guerrilla movements, but with different goals. Frelimo was Mozambique’s most potent armed group in the struggle for independence from the Portuguese during colonial rule. On the other hand, Renamo emerged after independence as a counter-insurgent anti-communist movement created and sponsored by the Rhodesian government to disrupt the Marxist-orientated and newly-formed post-colonial government of Frelimo. After the civil war and with the signing of the peace accords in 1992, Renamo made the transition to a political party. This coincided and was made possible by political reforms making Mozambique a representative multi-party democracy instead of the socialist one-party regime that characterized the previous two decades under Frelimo (Virtanen 2005). Frelimo, however, has remained the ruling party since the first elections in 1994.
Mozambique counts over two-hundred ethnicities and approximately forty officially recognised languages spread across the country’s massive 2700 kilometre north-south axis. This kind of diversity is comparatively high in sub-Saharan Africa. Yet the country has the interesting trait of being ethnically diverse but not ethnically polarised (Cheeseman & Ford 2007; Dowd & Driessen 2007). Inevitably ethnic identity is and will be socially salient in Mozambican society, but the question is to what extent this infiltrates its national politics and how it affects notions of Mozambican nationalism and state-legitimacy.
This thesis aims to test claims that stronger traditional systems (or pre-existing hierarchical social structures) will reduce the likelihood of ethnic politics in a country. Evidence compiled from quantitative and qualitative sources and analysis of sub-national variation suggest that the opposite argument is more accurate in Mozambique. This is likely because of the conflict between Frelimo and Renamo: modernisation versus tradition; centralisation versus decentralisation of political power. Traditional systems are demonstrably part of an ‘old’ Africa, where ethnic identity is more socially salient. As we will
5 see, Frelimo’s history of removing and banning traditional systems have likely reduced the salience of ethnicity in areas where its policies have ‘succeeded’. Where they have met fierce resistance, traditional systems remain strong. In addition, as we will see, ethnicity has more political salience than expected in Mozambique. I begin this thesis with an overview of the scholarly debate surrounding ethnic identity and political competition – particularly in sub-Saharan Africa – and what kind of indicators explain the variation in the strength of this relationship. Drawing on this literature, I will make a preliminary argument primarily based on the work and theories of Dominika Koter: that local intermediaries in Mozambique mitigate the political salience of ethnicity. I will test this theory through a qualitative case comparison of Mozambique’s regions, and to make the analysis more manageable, I will divide the country into a southern, central, and northern region. I will start with a consideration of the historical legacy of colonial and post-colonial politics in Mozambique and their effect on social structures. In addition, I have conducted an interview with a Mozambican expert – with extensive knowledge of the country’s history – as corroboration to my findings. After the historical analysis, I will compare the three regions with data from Afrobarometer. The next section will discuss why the claims with regards to the role of traditional leaders in electoral mobilisation and ethnic politics, might not apply in the case of Mozambique. Finally, I end with a conclusion and propose avenues for future research.
Literature Review: Ethnic Identity and Political Competition in Africa
& How Intermediaries Fit In
There is a large body of work that has focused on how politicians often try to appeal to voters based on their shared ethnic identity (Inter alia Posner 2005; Palmberg 1999; Horowitz 1985; Koter 2013, 2016). This ‘ethnic mechanism’ is often claimed to play a decisive role in shaping party politics in Africa and one of the primary cleavages that define the nature of politics on the continent. Traditionally, the most common explanation for this phenomenon is that voters expect to profit from a coethnic who is in power. As Van de Walle is cited saying in ‘King Makers: Local Leaders and Ethnic Politics in Africa’ (Koter 2013), “citizens may feel that only a member of their own ethnic group may end up defending the interest of the group as a whole and that voting for another ethnic group will certainly not do so” (van de Walle 2007, 65). Ethnicity in this case is assumed to be associated with certain interests and allegiances, which politicians use to their advantage. Voters might therefore be led to believe that a coethnic in power will lead to benefits, and will vote accordingly.Despite the plausibility of this narrative, the role ethnicity plays in elections is often more complex and its influence can vary per case. For example, there is variation in the strength of ethnicity as a predictor of party affiliation depending on whether we are looking at ruling or opposition political parties (Norris and Mattes 2003). Namely, opposition parties often exhibit more ethnic polarisation amongst its supporters than ruling parties – possibly the result of grievances of voters who feel ignored or have received no perceived benefits from the ruling parties. In Mozambique, we can see this to an extent as well: Renamo is slightly more polarised than Frelimo (Dowd & Driessen 2007). Similarly, Eiffert et al. (2010) contend that during the run-up to elections in Africa, the relative importance of ethnicity gradually increases.
Furthermore, the electoral system of a state can facilitate an increase or decrease in ethnic fractionalization of the political landscape (Colomer 2016, 431). One can end up with
6
votes dispersed over so many parties, no single one can achieve a majority; on the other hand, it is more likely in a case of high fractionalization, that the amount of parties will reduce and will try to appeal across different ethnic groups across both “inter-group and intra-group cleavages” (idem.). A similar study by Posner (2007) finds that one-party and multi-party regimes influence election strategies: multiparty elections will try to appeal more to broader groups (e.g. language or region) whereas one-party elections will have a stronger focus on “tribes or clans”. This would at first sight seem to apply to Mozambique as well, considering the multi-ethnicity of the parties and its system of multi-party rule.
Returning to the matter of voting for coethnics because of expected benefits, this might not always end up being reality. In fact, there is empirical evidence that shows how ethnic groups might even pay higher costs for having coethnics in power (Koter 2013). For example, in Kenya, the president’s ethnic group payed more taxes than other groups. The coethnics were probably more tolerant of unfair treatment from their president, possibly because having a coethnic in power brings more than merely financial and economic benefits. At the same time, it shows how one can end up with different conclusions depending on the variables you test. Either way, ethnic support is often based more on perceptions and hope than actual outcomes (idem.; Posner 2005).
Dominika Koter (2013) contends that many of these studies rely too much on quantitative data and mathematics, and lack an interest in the cultural dimension that might explain variation in the degree of ethnic politics in the various countries in Africa. Her research focuses on answering the question why ethnic electoral blocs form in some countries and not in others, despite seemingly large similarities between countries. Her theories suggest that the strength of traditional systems play a decisive role in determining the amount of ethnic politics in a country.
Theory: The Role of Local Intermediaries in Mitigating Ethnic Politics
I intend to test Dominika Koter’s theories (2013, 2016) about how powerful local elites can “[create] cross-ethnic allegiances and [avoid] ethnic voting blocs” (2013). Her work pays particular attention to Senegal and Benin, although West Africa in general forms the main area of her research. She writes about the essential role intermediaries can play during elections. These intermediaries are often local, mostly traditional leaders (such as chiefs) or otherwise religious leaders. In Mozambique, for example, these traditional leaders are collectively referred to as régulos (Virtanen 2005). These kind of figures have played important historical roles throughout Africa. Koter shows how the measure of influence local leaders hold in their area, affects politicians’ campaign strategies and how they appeal to the electorate. In her concluding remarks, she suggests her findings have the potential to be generalised across the continent and beyond. She explicitly states that one of the countries for which her research might be relevant is Mozambique (2013). This is one of the main reasons to test her theory and see if the cases in West Africa hold true in Mozambique. The role of traditional and other local leaders remains pertinent in much of Africa because so many states are characterised by weak governments and poor access to public goods. That void is filled in by local leaders. In exchange for material goods or other benefits, they are able to legitimize their role within the community. The relationship between locals and their leaders entails a process of give and take, where some form of exploitation by the traditional leaders can only be sustained if they perform adequately and provide enough benefits to their community in terms of access to resources. Economic or financial benefits7 are not always the only factor determining the authority and legitimacy of traditional leaders. Many if not most cultures have a kin-based system, often enshrining spiritual and symbolic power with the traditional leader. Hence locals may depend on their traditional leaders not only for economic benefits, but also cultural services and as a mediator in conflicts. Local leaders together create a network of intermediaries which potentially provides politicians with efficient and effective access to a large share of the electorate. When politicians make use of local leaders, she describes this as taking “advantage of pre-existing relations of dependence” (Koter 2013). On behalf of politicians or under their instruction, intermediaries can persuade their followers to vote for a particular candidate or political party. In addition, local leaders can strike a deal with politicians in return for some sort of payment – whether that is in cash, projects, or some other form of support can vary and is likely to depend on the needs of the community. If a local leader has authority in his community and his judgment is trusted, voters are more likely to follow his advice. The key is this: the politician they are to vote for does not necessarily have to be from the same ethnic group. Instead, what is more important is the belief that a candidate can deliver the goods. The connection between their local leader and a candidate – reducing the distance of communication – might reinforce that perception too. It is in the local leader’s interest to try to pursue some benefits for his dependents, as his legitimacy depends on it. The more important a local leader is for the wellbeing of his community, the more likely they are to listen and act according to his recommendations.
This exchange for political support is a feature of much of politics in Africa (Koter 2013), otherwise described as clientelism. Koter (idem.) describes it as the following, “Instead of reaching out to the public on the basis of impersonal ties, politicians can create constituencies through personal connection between voters and their leaders.” To politicians, appealing to voters with this strategy gives more confidence that there will be political support. The bond between voters and their community (and hence, local leaders) is much stronger than external networks. In addition, the larger the perceived distance between a voter and a political candidate, the harder it is to persuade voters effectively. Parties for example can hand out gifts before an election but there is no enforcing mechanism to ensure people vote accordingly (idem.). However, institutionalized parties might be able to have a tighter grip on proceedings because they can employ party agents to monitor throughout the country – this might, for example, be the case for a former one-party regime such as Frelimo. Delegating political mobilisation to intermediaries allows campaigners to much more effectively monitor the performance of these ‘constituents’ based on election results per area. It becomes much more clear which area votes what.
In summary, based on the mechanisms described here we expect a number of things in Mozambique. In regions where traditional systems are strong, the inverse should be true for the political salience of ethnicity. This is because traditional leaders – or perhaps more broadly, intermediaries – are able to appeal across ethnic lines, making it more rational for politicians to make use of such leaders. If these leaders are weak, we expect to find an increase in ethnic politics. The strength or weakness of local leaders is the result of historical trends and the legacy, in all likelihood, of a state’s colonial experiences. Colonial powers can either have uprooted or made use of (and possibly strengthened) social hierarchical ties. In addition, this mechanism is mostly relevant to rural areas.
8
Qualitative and Quantitative Evidence
For the analysis, Mozambique’s 10 provinces (and the capital Maputo) will be categorized into three regions for comparison: North, Centre, and South. The evidence will be presented in two sections. The first section is an analysis of the historical legacy of colonial and post-colonial politics on the traditional systems and relations of dependency in various regions of Mozambique. A number of anthropologists have conducted important research on post-civil war Mozambique, such as Lubkemann (2008), West & Kloeck-Jenson (1999), Christiaan Geffray (1989), Pekka Virtanen (2005). In addition to calling upon these sources, I have conducted my own interview with Carlos1, an experienced Mozambican civil worker and former member of Frelimo. Through him, I can corroborate and add to the evidence from researchers such as those mentioned above. The historical analysis is divided accordingly into four subsections: the colonial period, the emergence of Frelimo, the start of the postcolonial period, the emergence of Renamo and on-set of Civil war, and finally the post-civil war period. In the second section we will refer to quantitative data from the Afrobarometer round 6 survey, published in 2016, to corroborate the evidence from the historical section. We will be looking at strength of local leaders and the salience of ethnicity in Mozambican society, as well as what the national election results tell us about the political culture in Mozambique. The data was collected in 2015 shortly after the 2014 national elections. Although the survey has a sample of 2400 respondents, recoding in SPSS means we will draw from a maximum of 2180 respondents. Furthermore, the population per province and rural-urban proportions are taken into account in the share of respondents per province. The official results from the national elections of 2014 is provided by the Mozambique Bulletin Board.Historical Legacy
Colonial Period When the Portuguese arrived in what was yet to be called Mozambique, they implemented an administrative structure for their newly established colonial state. Each administrative post was called a regedorias, roughly translated as ‘chiefdom’. The Portuguese would appoint locals to run these posts, and these positions were called autoridades gentillicas. The highest ranking among these autoridades were called régulos. The main tasks of these intermediaries were to collect taxes, recruit labour, and enforce colonial rules amongst other things. According to West & Kloeck-Jenson (1999), the Portuguese would take into account the pre-existing hierarchies and “attempted – when possible – to incorporate them … into the colonial state’s own system of native administration.” However, it was not uncommon to appoint locals who were not ‘traditionally’ recognised as leaders prior to the arrival of the Portuguese. This could be because an area was previously unoccupied, not under some form of central rule, but could also be because the Portuguese thought giving power to ordinary locals might make them more loyal.How the population experienced the arrival of the Portuguese and the implementation of their administrative structure, varied per region. In the south – in modern
1 Carlos is not his real name. He has asked to remain anonymous for professional and security reasons.
9 Gaza and Inhambane – the colonials imposed a hierarchy of autoridades gentilicas that had three levels, similar to the hierarchical structure that was already in place prior to the arrival of the Portuguese in the area (West & Kloeck-Jenson 1999). In the north, the Portuguese encountered a tiered system for extracting tribute instated by slave-lords, chiefs, and kings who had enslaved people amongst their own as well as amongst their neighbours. Forms of tribute, coercion and state-like organization were already commonplace in some areas. Suggestions therefore that the European colonists brought with them an alien form of administration that destroyed and uprooted local systems, is not entirely accurate. In many cases the Portuguese made use of existing systems but the tasks these intermediaries carried out might have differed qualitatively to the kind of roles they traditionally carried out in their communities. For example, the focus could be on policing and recruiting workers rather than mediating in disputes or overseeing marriages. Traditional systems were therefore already affected and transformed by conquerors preceding the arrival of the Portuguese. At the same time the legitimacy of these local leaders in the long run would depend on their performance. If the local population was dissatisfied, they could be dismissed by a higher authority and replaced by another local. However, the passing on of titles after death or retirement often remained kin-based in many areas, a concept familiar and recognized throughout Mozambique historically (idem.).
Emergence of Frelimo
The Mozambican independence movement started to organize in Tanzania in the late 60s, and later became known as Frelimo (Liberation Front of Mozambique). The leadership of the movement was a unification of traditional rural leaders and educated urban elites (Lubkemann 114). As Carlos, my Mozambican contact told me, the interests of people within the movement were initially more orientated towards ethnicity and tribal groups. Each group was more concerned with their ‘own’ problems with respect to the rule of the Portuguese rather than the abstract ideal of creating ‘one’ Mozambique, as the pan-Africanists idealised (Interview I). However, the colonial power was a common enemy and thus made it possible for people who had previously rarely interacted with each other –and had opposing views on many issues – to lock arms and establish the Frelimo movement. The rural leaders were more motivated by their experience of harsh policies from the Portuguese whereas the educated elite were mostly inspired by the pan-Africanism of figures such as Kwame Nkrumah (Interview I; Lubkeman 2008: 114). The Portuguese had instated agricultural policies that forced locals to grow only cash-crops and were forbidden from growing food for sustenance; this lead to severe food shortages and starvation in northern regions. As a consequence, grievances and resentment developed amongst the rural population towards the Portuguese rulers. These grievances made it possible for different ethno-lingual groups, who until then had limited interaction with each other, from uniting against a common oppressor. As Eduardo Mondlane – the first president of Frelimo – wrote, “Nationalism, like in all African nations, is born out of the oppression of the colonists” (Mondlane 1964).
There were tensions between the rural and urban elite and these finally came to a head with the assassination of Eduardo Mondlane in 1969. The more educated urban elite were proponents of “radical social revolution” (Lubkemann 2008: 115) inspired by pan-africanism and communism while the rural elites had their reservations about changing traditional systems, particularly the concept of kin-based leadership. The perpetrators or exact motivations behind the assassination are shrouded in mystery, but rural elites
10
abandoned the movement and from then on Frelimo was consolidated by the urban elite and became a more overt Marxist-Socialist movement. That key leaders from the rural elite “defected to the Portuguese” reinforced suspicions amongst the Frelimo leadership that traditional leaders were responsible for the murder of Mondlane and facilitated the Portuguese colonial administration. The rural leaders that did not defect tried to organize a rival organization based on ethnicity. (Lubkemann 2008: 114) Frelimo’s military activities remained mostly confined to pockets of areas in the far north of the country, along the border of Tanzania. Where it was able to capture (or ‘liberate’) and control territory, Frelimo started to organise villages (moving people from the fields to a more centralised organisation) and successfully imposing its policies and visions of a ‘modern’ and new Mozambique amongst the ‘liberated’ population (Lubkemann 2008: 141). This meant challenging and replacing traditional authorities, who in the eyes of Frelimo, were considered illegitimate and pawns of the colonial regime (Lubkemann 2008: 115). In the eyes of Frelimo, any semblance of the old Africa had disappeared with the arrival of the Europeans – present ‘traditional authorities’ were simply an illusion and a fallacy because, according to Frelimo, they had been appointed by the colonists. Postcolonial government A coup d’état in Portugal in 1975 led to the sudden de-colonization of Mozambique. Frelimo was given the reigns to the country by the new Portuguese government. The new post-colonial government acted swiftly and started to implement large-scale radical social policies to ‘re-educate’ the population and enforce the state’s authority.
One of the most crucial (and destructive) Frelimo policies, was to ban traditional systems and remove traditional leaders from power and delegitimize the legal authority they had under the colonial regime. The colonial system of regedorias was replaced by grupos
dinamizadores – the smallest unit of governance of the state – which were headed by secretariós, similar in status to régulos under the Portuguese (Lubkemann 2008: 119). The
legitimacy and power of the secretariós versus that of the régulos in the eyes of the population would become a central feature in the build-up to the civil war in Mozambique. Frelimo’s measures had an enormous social impact in the country, particularly in rural areas. People previously recognized by the population as traditional leaders, embodied with symbolic and spiritual significance, as well as religious leaders, were suddenly delegitimized by the central government and excluded from any leadership roles (Lubkemann 2008: 119). Instead new leaders – secretarios, namely – were instated that were often drawn from marginal groups within a community and did not fulfil any of the criteria locals by which they legitimized local or traditional authority (idem.). These kind of acts struck at the “core tenets of local authority around which community governance had been arranged” for nearly a century and in a far more traumatic, “dramatic and profound way” than the “occasional interference of Portuguese officials” or the introduction of Christian religion by missionaries (idem.: 122; Interview I) “Many vividly recalled” traditional institutions of the regúlos and other traditional leaders being publicly debased (idem.: 121). This and the manner in which Frelimo went about ‘re-educating’ the population was often described as “profoundly humiliating” (idem.: 123).
The key to this story is that each region responded differently to Frelimo’s policies based on their previous experience of the colonial regime and the liberation movement. This also affected what people expected under the new regime and how they interpreted
11 independence. As Lubkemann (2008: 117) shows through his conversations with locals in the central province of Sofala, their experience of the colonial regime was less intensive than, for example, in the northern regions of the country (Tete, Niassa, and Cabo Delgado), where people had experienced harsher treatment from the Portuguese and had witnessed armed rebellion against the colonial regime. Nor had central regions been as affected by forced villagization by the Portuguese. For the more central regions, independence came “virtually [as] an overnight event” (idem.: 118). Coupled with that, the region had a long, historical distrust and dislike for centralized government (possibly because the ‘central’ power felt distant). This was reinforced during the colonial period but even more so with the post-colonial government.
The state-building process Frelimo had started – which was not only confined to establishing local administrators and leaders – led to instability in the country. This ultimately had severe consequences and contributed to a violent response in the form of a civil war. Frelimo retained negative views of traditional systems. For example, during the liberation period, Frelimo had to contend with ethnic animosities in northern regions as they tried to expand into the Makua areas – predominantly in Nampula – as well in Niassa and Cabo Delgado, which border the north of the province (idem: 119).
Emergence of Renamo and Civil War on-set
In 1990, Geffray was one of the first researchers to publish a major (anthropological) study on the civil war in Mozambique. He argued that Frelimo’s “disenfranchisement and humiliation” (cited in Lubkemann 2008: 153) of traditional authorities and the suppression of traditional identity was the catalyst for popular support for Renamo, a militant group created and sponsored by the Rhodesians in response to Mozambique’s independence. Renamo made use of this growing resentment towards Frelimo. Lubkemann reports how Renamo agents would promote Renamo’s agenda as “getting rid of the secretariós, re-establishing the old régulos…dispersing the communal villages [aka as aldeias communais, forced villagization policy by Frelimo]” (idem.: 144) amongst other Frelimo policies to be reversed or challenged. In addition, these agents would inform and make use of local leaders to raise support amongst their community.
Lubkemann questions to what extent Renamo’s “hearts and minds” campaign was genuine. The jury is still out to what extent Renamo had a real political vision, even in its present form as a political party. Instead, Lubkemann argues it is rather the perceptions people had of Renamo as an anti-Frelimo party and a movement that would reduce the interference of a central government in local affairs and customs that made people choose to support them (Lubkemann: 138, 143). People felt their “very own social existence [was being negated by Frelimo]” (Geffray 1989: 19).
In particularly central – but likewise in northern provinces such as Nampula – traditional systems were strongest. This largely correlated with the fact these societies had very dispersed settlements. In contrast, the northern provinces had experienced forced villagization by the Portuguese as well as forced relocation under the direction of Frelimo, weakening even further already fragile traditional systems as people started to live in more concentrated settlements (Lubkemann 2008: 142). It is therefore no surprise to find that popular resistance against Frelimo’s policies were stronger in the provinces with large rural populations and more dispersed patterns of settlement. According to Geffray, Frelimo’s failure was to ignore the variation in social structures and their strength throughout the
12 country, particularly in the central regions (Geffray 1989). Frelimo’s policies were intrusive, radical and disrupting to the norms in central regions, fostering fierce resistance and causing grievances towards the post-colonial government. As violence increased throughout the country, trends along ethnic lines could be seen in the country. This specifically relates to the Ndau and Shangaan ethnic groups – the former largely based in Sofala and Manica, and the latter in bordering Gaza, to the south. The Shangaan are often perceived to have a sense of superiority over other ethnic groups, and as an outside ethnic group, moving through Shangaan areas would often mean entering subordinate relations with the Shangaan (Lubkemann 2008: 182). My Mozambican contact confirmed these contentious relations, as well as corroborating Lubkemann’s comments that this sentiment applies to most groups of central and northern origin in relation to the Shangaan. Practically what this meant, was that in the run-up to and during the civil war the Ndau were associated with Renamo and the Shangaan with Frelimo. The Ndau would flee into Renamo held areas and the Shangaan into Frelimo enclosures (idem.: 134). This could inevitably have affected how these groups retained or lost their connection with traditional systems. Worth mentioning as well, is that Frelimo’s leadership was dominated by southerners, as were many local officials and secretariós installed in more central and northern districts (idem. +carlos).
Since 1992 Peace Accords
Since the end of the civil war and the establishment of peace, Frelimo’s position towards traditional authorities has become more nuanced. The first post-war president of Mozambique, Frelimo’s Chissano, is on record saying “we want traditional authority to exist” (cited in West & Kloeck-Jenson 1991). Instead of undermining traditional authorities – having probably come to the realization its policies were counterproductive – Frelimo have apparently started to make increased use of traditional systems during elections. In fact, the official status of traditional and local leaders remains a tricky talking point in present-day Mozambique. According to West & Kloeck-Jenson, traditional leaders remain a particular potent and “salient political force” in rural areas. Furthermore, it seems that the Frelimo leadership – as Chissano’s comments suggest – realised that it is in their best interests to “strike a peace” with traditional authorities where they weren’t able to remove or make allies of them (idem.). A particular section in West and Kloeck-Jenson’s research recounts an interesting case of a village called Cariua, where the position of ‘traditional authority’ kept shifing to another person depending on who established military control in the area. The case goes as follows: in a locality within the central province Zambezia, a founder established a settlement and took on a position as chief. His heirs retained his name and title in the settlement until the Portuguese arrived. They replaced the chief with someone else as a régulo, an outsider, in the hope that they would have a more loyal representative. After independence – and with the banning of traditional authorities – Frelimo chose for another individual not related to the original founders or subsequent chiefs, as secretario of the settlement. However, during the civil war, Renamo gained control of the settlement. The residents masked the identity of their leader who remained amongst them and told the Renamo fighters the secretario had fled. Renamo then appointed a young nephew of the original founder (who was still alive) as chief of the settlement. When the original founder and new chief were captured after an attack from Frelimo, the villagers ironically reinstated the secretario but now as a ‘chief’,
13 unbeknownst to Renamo. When the war ended he was reinstated to his original position, as secretario. However, he had to work side-by-side with the original régulo from then on as well, who had returned from captivity.
In summary, the Portuguese implemented an administrative structure that often adapted existing social structures in the various regions of Mozambique. In some areas, where forms of centralised rule were non-existent or hierarchical structures weren’t very layered, the Portuguese created and appointed new ‘traditional’ leaders. Frelimo’s influence on the country was even more profound by literally banning traditional systems and excluding former traditional leaders from any leadership roles. In some areas, Frelimo was able to succeed re-engineering communities in their vision: removing traditional systems from society. However, in large swaths of the centre and in parts of the north of the country, the movement encountered fierce resistance. Here traditional systems were at their strongest and not amenable to Frelimo’s radical policies. The resentment towards Frelimo was capitalised upon by the externally sponsored counter-insurgent movement Renamo, ultimately leading to a long, devastating civil war. Since the peace accords in 1992 there is a strong divide between particularly the centre of the country and the rest, where in the former traditional systems still wield considerable power, areas held by Frelimo in the north and south have considerably weaker traditional systems, if any of it is left. Nevertheless, Frelimo has softened its stance towards traditional authorities and realizes it has to use them to maintain its power as the ruling party of the country.
14
Survey Data Comparison
Mozambique is organized according to the following ten provinces and the capital Maputo, in geographical order from south to north: Maputo City, Maputo Province, Gaza, Inhambane, Sofala, Manica, Tete, Zambézia, Nampula, Niassa, and Cabo Delgado. For the purposes of the test, the provinces have been categorized into three regions: South, Centre, and North. The provinces are divided as follows: Maputo, Gaza, and Inhambane are part of the South; the provinces Sofala, Manica, Tete and Zambézia are part of the Centre; the remaining provinces are part of the North. The decision to arrange the provinces this way is based on the clear ethno-lingual and cultural differences between the three regions. Each region is made up of multiple ethnic groups but the languages are similar enough that everyone can communicate with each other; this is not the case when it comes to the communication between the regions. As a result, these three regions have distinct cultures and histories each may be considered as a political unit (Interview I). In addition, this allows a more manageable case comparison and affords us a larger sample per case than if we were to take each province as the largest unit. Figure 1: Left, Presidential Elections; Right, Parliamentary Elections 1. NATIONAL ELECTIONS 2014 Region President Parliament Nyusi(Frelimo) Dhlakama (Renamo) Frelimo Renamo
South 69% 21% 64% 20%
Centre 48% 48% 49% 46%
North 45% 49% 46% 45%
15
The results of the national elections in 2014 are reported in Table 1 and visualised per province in the accompanying Figure 1. What we see is that Frelimo dominates particularly in the South and has some influence in the two most Northern states of Niassa and Cabo Delgado. The two central states of Manica and Tete are close-fought: in the presidential elections Renamo wins whereas for the parliamentary elections Frelimo is the victor. All other states return the same winner in both the parliamentary and presidential elections. The region-by-region breakdown of the national elections can be found in Appendix 1. For Cabo Delgado you will see an oddity as people overwhelmingly – nearly eighty-percent – voted for President Nyusi; it so happens he is from the Makonde tribe, who originate from this province. Overall, Frelimo receives nearly sixty-percent of the votes in the election, despite the fact the Centre (the largest region by population) and the North show a slight preference for Renamo. In Table 2 we find the results for Party Affiliation of the respondents in the Afrobarometer survey. These results are based on question 99 (Who would you vote for if the election were tomorrow). Another question that could have been used as a proxy was question 90, which asked about party affiliation directly. However, the question was asked in two stages (90a + 90b) creating difficulties for recoding in SPSS. Question 99 and 90 had nearly identical results, so the choice was made to use the former as a proxy for party affiliation. The first thing that one notices is that very few of the respondents would vote for, or felt affiliated with, Renamo: a mere 15% compared to the approximate 35% during the national elections. The total result of 56% for Frelimo is in accordance with the official results, although there is some deviation in the subnational variation.
The reasons for the inaccurate results can be manifold. Afrobarometer’s method might contain a systematic error, avoiding ‘Renamo’ areas. However, there are enough sources that describe a culture of fear in Mozambique (Virtanen 2005). In addition, in my interview with my Mozambican contact, this element kept returning: “People who vote Renamo are will often not admit it in public…” (Interview I). In the Afrobarometer data, 26% of respondents refused to answer the question regarding their party affiliation. The figure was highest in the Centre, where Renamo considered to be most popular.
I have included three more variables based on the Afrobarometer survey to find evidence that respondents might have refused to answer because they did not feel comfortable admitting their preferences for Renamo. The theory is that Frelimo can have a tight grip in the region. The first is the question 49 regarding the respondent’s fear for political intimidation or violence. As we move towards the north of the country this fear increases. In addition, at least two-thirds of the respondents in the central and northern provinces
16
admitted to having such fears. The second variable is based on the judgment of the interviewer to what extent a respondent had difficulty answering the questions. This seemed particularly the case in the central regions. Lastly, I looked at the amount of times local representatives or party officials interfered with the survey conducting its interviews. The incidence rate is very small in the south and the north, but there is certainly a much higher incidence rate in the central regions.
The three variables together suggest that fear might have a strong effect on the honesty and openness of the respondents in the survey. In combination with the literature and my interview with Carlos, the decision was made to recode the variable such that the significant majority of those who ‘Refused to Answer’ are assumed to be ‘Renamo’ supporters. To increase the size of the sample, we have also categorized MDM as Renamo, as the former is a non-militant off-shoot of the latter and will appeal to the same voter base (Virtanen 2005). These decisions could be considered scientifically irresponsible, as not all those who refused to answer will be Renamo supporters. Nevertheless, the results will likely be more accurate and reflective of people’s preference in Mozambique. 3. PARTY AFFILIATION RECODED Region Party Frelimo Renamo (+ MDM) South 64% 36% Centre 41% 59% North 67% 33% Total 56% 44% Table 3 shows the result of recoding question 99 to include MDM voters and those who refused to answer as de facto Renamo supporters. We find that this action certainly made the total election results more similar to the official national election results. In addition, about six-percent of the vote went to MDM, which would suggest – if this figure is subtracted from Table 3’s total – around 38% of the electorate voted for Renamo, which is very close to the 34-36% registered in the national election of 2014. However, the results per region are more clearly either in the direction of Renamo or Frelimo. However, these figures don’t reflect voting but affiliation; some people might not vote for their preferred party because of fear or there might be cases of tampering or voter fraud. As Carlos tells me, he once witnessed a Frelimo observer take a voting box and light it on fire.
Testing the Ethnic Mechanism
Before we consider the role of intermediaries in political mobilisation in Mozambique, we will look at the importance of the identity of the political parties or its leaders according to respondents. In Table 4 (pg. 17), three variables (Question #62) reflect what respondents feel are the main differences between the political parties or their leaders. As we move towards the north of the country – away from the capital city, Maputo – the ethnicity of leaders increases in importance. In the Centre, twice as many respondents indicated that the religion of the leaders was the most important difference.
17
18
All three regions regard the regional identity of the politicians with the same importance. These three variables considered together – around twenty-percent of the Centre, seventeen in the North, and eleven in the South – we find that the regions have different views about what the main differences are. However, overall, respondents place more importance on other aspects such as the integrity of the leaders or the economic/development programmes of the candidates. Nonetheless, these figures represent a minimum. The question is formulated such that respondents have to identify the ‘main’ differences between leaders and parties. Hence, we can assume that a larger proportion of respondents will consider ethnic, religious, or regional identity important considerations. In Table 5 we look at two further variables: how ethnic groups are treated and what kind of nationalism they identify with most. Half of the respondents in the Centre and the North felt that their group was treated unfairly at least occasionally, whereas this figure is a mere fifteen-percent in the South. The question here is to what extent unfair treatment overlaps with cases of discrimination. Either way, it suggests ethnicity is much more politically salient in the Centre and North than expected. It might be related to tension, for example, with the politically powerful Shangaan from the South as described in the previous section. The reasons for this can vary. In addition, my Mozambican contact pointed out during our interview that people in the North often feel aggrieved because they perceive most of the fruits of their labour is transported to the South; most of Mozambique’s economic activity in terms of value takes place in the North (Interview I).
The second variable looks at the degree to which respondents identify with their ethnicity compared to the idea of being a ‘Mozambican’. We find that people in the South feel mostly Mozambican, but as we move towards the north of the country – away from the capital – people feel more increasingly more connected with their ethnic identity. Again, there is mostly a strong divide between the South and the Centre & North. In the Centre twice as many respondents felt more connected with the notion of their identity than the idea of ‘Mozambican’, compared to the North where a larger proportion of respondents felt equally Mozambican as the ‘ethnic identity’. Nonetheless, the vast majority of respondents felt more Mozambican than anything else. Intermediaries In Table 6 we look at seven variables that consider the roles of intermediaries and (local) political leaders. For traditional leaders I have chosen three variables as a proxy: how often respondents interact with traditional leaders (contact); how much respondents trust their traditional leaders (trust); and whether they approve of their leader based on his performance the past twelve months. We find that the role of traditional leaders is most active particularly in the Centre but also in the North.
In the South, there is much less interaction with traditional leaders. However, the North has a lot more trust in their traditional leaders than in the rest of the country. Despite having more contact with traditional leaders in the Centre, respondents only trust their traditional leaders slightly more compared to people in the South. Perhaps this indicates the continuing struggle between secretarios and régulos – as what is a ‘traditional leader’ may depend on the interpretation of the respondent. This indicates that traditional systems are much stronger outside the South of the country.
19
Even more so, the approval ratings are almost as high as they can be at ninety-percent. Religious leaders play at least just an important role in the lives of people compared to traditional leaders. In this case, they also play a particularly stronger role in the South, and slightly more in the Centre. Religious leaders are interacted as often in North as compared to traditional. What certainly is more apparent, is that respondents have more trust in their religious leaders than in their traditional leaders. This is perhaps understandable, certain moral behaviour is more demanded of and associated with religious leaders than with traditional leaders. A third category to be looked at are political leaders at the local level. In this case I have selected three variables available to me as proxies: how often party officials or local government councillors are contacted; and how much people trust their local government council. Here we have it is difficult to ascertain a pattern. The Centre and North make use more often of party officials than in the South. When it comes to contacting local government councillors, the numbers don’t differ much between the regions, and in any case, are low. However, the trust in local councils are high in all regions. This is particularly the case in the north, where respondents almost universally indicate that they trust their local council. The figure is lowest at seventy-three-percent in the Centre. The variation is hard to interpret here, but respondents certainly place more importance in their religious and traditional leaders.
Analysis & Discussion
The results suggest that the southern regions of Mozambique are quite different to the central and northern regions in the political salience of ethnicity as well as the strength and relevance of traditional authorities. Our initial argument was based on Koter’s theories about the role intermediaries could have in affecting the degree of ethnic politics in a country. The expectation was that in countries where reliance and trust in local leaders were comparatively high, the inverse would be true for the degree of ethnic politics. For much of Mozambique, the reality seems a lot more complicated. In the South, traditional systems are weakest yet ethnicity seems least relevant on a political level. In fact, where traditional authorities are strongest and most trusted, ethnicity is more important and respondents indicate substantially higher rates of unfair treatment. Admittedly, Koter’s research is more robust and additional field research would have to be carried out to corroborate our findings. Despite the fact the Centre and North have similar scores for the various indicators, the Centre showed a preference for Renamo whereas the North preferred Frelimo by a large margin. This would suggest that the strength of traditional leaders has no influence on the popularity of one party over another across the country. However, this does not align with the data and historical records in previous studies. Moreover, the official election results suggest Renamo should be more popular in the region in reality. But as West & Kloeck-Jenson (1999) demonstrated in their account of the Cariua locality, the definition of who or what is a ‘traditional leader’ can change rapidly over time, particularly in the most northern provinces. These might irreparably skew the results. The reality is we do not know based on current evidence, but it could be that because ‘traditional authority’ is a fluid term that the data from these questions cannot be generalized – not within countries, let alone between countries.
In addition, although there is much written about the significance of traditional leaders in Mozambique, research has focused mostly on their involvement in the war rather than how they currently mobilise supporters for elections. West & Kloeck-Jenson certainly allude to the fact that they play a relevant role in elections, but how exactly and what the mechanisms are remain unclear. The question of how politicians can persuade traditional
20 leaders to support them – particularly Frelimo, who have done so much to undermine them in the past – remains pertinent and unanswered. Nonetheless, Mozambique might prove as a counterpoint to Koter’s research due to the unique history of the country compared to the rest of the continent. The issue of the legitimacy of traditional authorities played a central role in the schisms that tore the country apart and it continues to be a difficult topic in Mozambican national politics. Much of the country’s conflicts can be summarised as a struggle between the old Mozambique and a vision of a new one; or in other words, opposing ideas about how the state achieves legitimacy. Invariably, because strength of traditional systems varies per region and per ethnic group, this correlates with support for or against Frelimo. The stronger the traditional system, the more likely there is resistance towards Frelimo. There seems to be a strong conflict between centralised and decentralised rule in the country, which means that the legitimacy of the traditional leaders themselves become a political issue. As a result, because some regions place more importance on traditional systems and each region is populated by a particular ethnic group, one might come under the impression that there is a stronger relationship between ethnicity, traditional leaders, and political affiliation. My Mozambican contact argues that there is a potential for ethnic conflict in the country, but that the issue remains dormant (Interview I). The civil war is still too fresh a memory (“no one wants to return to those times”) and, for example, the South needs the North economically and so Frelimo will take every measure possible to keep the country together.
Conclusion
Although local leaders in Mozambique might act as intermediaries during elections to mobilise voters, there is not enough evidence to suggest they also reduce ethnic politics in the country. In fact, the test results suggest the opposite: where traditional leaders are more active, ethnic identity appears to become more important. This is likely explained by Mozambique’s history and how it is defined by the struggle between a movement that seeks centralised rule and an opposition that wants decentralised rule. Decades of oppression and radical social policies have created deep resentment towards Frelimo amongst rural populations in the central regions. Because the southern, central, and northern regions roughly act as proxies for ethnic groups, there appears to emerge a correlation with the strength of traditional systems. Strictly looking at the results from the Afrobarometer data, strong or weak traditional systems do not swing the election in favour of one party over another.
One of the key questions that remain unanswered, is how local leaders mobilise their dependents and how these leaders interact with the two main parties in the country. In addition, an interesting question to pursue is how secretarios and (former) régulos exercise their influence over their constituents. Do the secretarios work against the régulos, or do they interact with them to try and persuade them to support Frelimo? Either way, to uncover such answers it is essential to undertake an independent field research. The Afrobarometer proved a useful addition in the analysis, but the multi-interpretability of some of the questions for the respondents makes it on occasion difficult to say with certainly what exactly we are measuring.
As a concluding remark, the fact this case might not align with claims about the potential mitigating role of traditional leaders in ethnic politics, is precisely because the very legitimacy of traditional leaders in Mozambique is at the centre of so many of the country’s conflicts and political problems.
21
22
References
Cheeseman, N., & Ford, R. 2007. Ethnicity as a political cleavage. Afrobarometer Working Paper No. 83. Colomer, J. 2016. The Handbook of Electoral System Choice. Palgrave Macmillan Press. Dowd, R.A. and Driessen, M., 2007. Ethnically Dominated Party Systems and the Quality of Democracy: Evidence from Sub-Saharan Africa, Afrobarometer Paper No. 92. Eifert, B., Miguel, E. and Posner, D.N., 2010. Political competition and ethnic identification in Africa. American Journal of Political Science, 54(2), pp.494-510.Geffray, C., 1990. La cause des armes au Mozambique: anthropologie d'une guerre civile. Karthala Editions. Horowitz, D.L., 1985. Ethnic groups in conflict. University of California Press. Koter, D., 2013. King makers: Local leaders and ethnic politics in Africa. World Politics, 65(02), pp.187-232. Koter, D., 2016. Beyond Ethnic Politics in Africa. Cambridge University Press. Kyed, H.M. and Buur, L., 2006. New sites of citizenship: recognition of traditional authority and group-based citizenship in Mozambique. Journal of Southern African Studies,
32(3), pp.563-581.
Lubkemann, S.C., 2010. Culture in chaos: An anthropology of the social condition in war. University of Chicago Press.
Mondlane, E., 1964. The development of nationalism in Mozambique. Dec, 3, p.8.
Norris, P. and Mattes, R., 2003. Does ethnicity determine support for the governing party? The structural and attitudinal basis of partisan identification in 12 African nations. Palmberg, M., 1999. National identity and democracy in Africa (Vol. 95). Nordic Africa
Institute.
Posner, D.N., 2005. Institutions and ethnic politics in Africa. Cambridge University Press. Posner, D.N., 2007. Regime change and ethnic cleavages in Africa. Comparative Political
Studies, 40(11), pp.1302-1327.
Van de Walle, N., 2007. Meet the new boss, same as the old boss? The evolution of political clientelism in Africa. Patrons, clients and policies: Patterns of democratic
accountability and political competition, pp.50-67.
Virtanen, P., 2005. Tradition, custom, and otherness: The politics of identity in Mozambique.
Identities: Global Studies in Culture and Power, 12(2), pp.223-248.
West, H.G. and Kloeck-Jenson, S., 1999. Betwixt and between:‘Traditional authority’and democratic decentralization in post-war Mozambique. African Affairs, 98(393), pp.455-484.
23
Appendix I
National Election Data from Mozambique Bulletin Board, edited by Joseph Hanlon and published by CPI and AWEPA
PRESIDENTIAL ELECTIONS 2014
RENAMO FRELIMO MDM total
Niassa 111114 120818 17766 249698 Cabo Delgado 75768 324857 15729 416354 Nampula 375592 339143 44587 759322 Tete 231972 218850 20493 471315 Zambezia 327300 268856 56983 653139 Sofala 229341 146043 36189 411573 Manica 169359 167912 13182 350453 Inhambane 57523 228819 15920 302262 Gaza 11591 342470 11098 365159 Maputo 71407 298624 35543 405574 Maputo Cidade 86326 287674 43868 417868 TOTAL 36% 57% 6% PARLIAMENTARY ELECTIONS 2014
MDM RENAMO FRELIMO TOTAL
Niassa 18105 91743 113496 223344 Cabo Delgado 19175 69167 310608 398950 Nampula 57416 299311 306676 663403 Tete 26343 197922 208933 433198 Zambezia 65033 283036 243096 591165 Sofala 53040 182049 138493 373582 Manica 16265 147886 157835 321986 Inhambane 16139 39795 201620 257554 Gaza 10454 7249 269914 287617 Maputo 47998 66684 265508 380190 Maputo Cidade 64490 82447 257829 404766 TOTAL 9% 34% 57% 4335755
24
Appendix II
Questions for Interview 1. In your experience, does ethnicity play a role in voter preference or party affiliation in Mozambique? (Yes/No) 2. If yes, which ethnic groups? 3. Is there sub-national variation, as in: is there a case of ethnic voting in some regions of Mozambique but not in others? 4. Do traditional authorities play a role in political mobilisation - working as intermediaries for FRELIMO or RENAMO to win votes? 5. Does the importance of traditional authorities vary per region or ethnic tribe?6. Will RENAMO voters not openly support the party? In a survey carried out by Afrobarometer, party affiliation of Mozambicans was heavily skewed towards FRELIMO (80% vs 16% for RENAMO). In addition, about 50% refused to answer the question. Again, is there a fear to be honest about party affiliation in Mozambique?
7. I would like to know how FRELIMO and RENAMO compete for votes in the central states of Mozambique (Sofala, Zambezia, Nampula) which seem to be hotly contested and have a slight preference for RENAMO. What kind of strategies do the parties use to win votes? Do they use the aforementioned intermediaries? Along which political cleavage do they appeal to voters? Instead of ethnicity, is it for example the urban elite (FRELIMO) versus the rural population?
