INTRODUCTION
The long-standing debate about the proper role for Africa's traditional leaders in contemporary politics has intensified in the last two decades, as efforts to foster democratisation and decentralisation have brought competing claims to power and legitimacy to the fore, especially at the local level. Not only have Africa's traditional leaders survived – often against the wishes, decrees and predictions of their detractors – but in many places they have actually thrived. Yet African governments continue to struggle with how to interact with these leaders, a challenge that becomes particularly acute in those countries that have extended democratic reforms to the local level. Analysts, too, remain uncertain about how the persistence of the chieftaincy will affect the prospects for democratic consolidation.
For years the discussion was dominated by a polarised debate between ‘modernists’ and ‘traditionalists’ (Keulder Reference Keulder1998), or ‘trivialisers’ and ‘romanticisers’ of traditional authority and institutions (Oomen Reference Oomen2000: 16), that was unrealistic and often unhelpful. But whatever their strengths and weaknesses, the continuing role and influence of traditional leaders is now widely accepted as a given. While there are exceptions, in most countries traditional authority does not by any means appear to be in the process of ‘withering away’ as the modernists once both hoped and predicted. It thus becomes essential to account for the role of traditional leadership and institutions as part of, or alongside, democratisation processes. This is so not because these authorities are in some way inherently good and democratic (as traditionalists once argued), but because, for better or worse, they are there, and they are clearly important exercisers of ‘public authority’ on the ground in much of Africa (Lund Reference Lund2006). As a result, in the last decade the discussion has increasingly shifted to a more pragmatic focus on understanding why these institutions persist, how they have evolved and are continuing to adapt to their changing political environments, and how various governments are coming to terms with them.
But the pragmatic shift in the direction of the debate cannot obscure the fact that there are still fundamental disagreements about whether traditional authority and democratic governance are ultimately compatible or contradictory, and about whether the two can be blended in viable and effective ways. Does cooperating with chiefs put at risk the very principles of democracy? Is democracy threatened, for example, if elected leaders must compete head-to-head with selected chiefs in what many see as a zero-sum game for control over power and resources? Can the loyal subject of a chief also be a committed democrat? Does engaging with the still patriarchal chieftaincy necessarily undermine the rights and freedoms of women? Does the persistence of traditional systems present a fundamental challenge to the very possibility of consolidating democracy on the continent? Or, on the other hand, is the chieftaincy a potential resource, an asset that can enhance the understanding, acceptance and effectiveness of democratic institutions under the right circumstances?
Survey data collected in Rounds 1 (1999–2001) and 2 (2002–3) of the Afrobarometer offer new insights and a more concrete basis on which to evaluate the merits of these claims. The results from over 35,000 face-to-face interviews in up to fifteen countries can help us to better understand popular perceptions of traditional leaders, how they are formed, and how they relate both to perceptions of elected leaders and to support for a democratic system of government.
Our data indicate that Africans who live under these dual systems of authority do not draw as sharp a distinction between hereditary chiefs and elected local government officials as most analysts would expect. In fact, far from being in competition for the public's regard, traditional leaders and elected leaders are seen by the public as two sides of the same coin. Overall, popular perceptions of traditional leaders are slightly more positive than those of elected leaders. But popular evaluations of both traditional leaders and elected leaders are strongly and positively linked. They appear to be consistently shaped by each individual's ‘leadership affect’, and by an understanding of chiefs and elected officials as common players in a single, integrated political system, rather than as opponents in a sharply bifurcated one. Thus, positive perceptions of chiefs go hand-in-hand with positive assessments of elected leaders, and vice versa. This connection is especially strong between traditional authorities and local government leaders. Far from fighting a pitched battle for public support, the fates of each appear to be inextricably linked. In contrast, an individual's level of modernisation plays a much smaller role in shaping perceptions of traditional authority than we might have expected. And most significantly for the debate about how democratic traditional rule is, there is no evident conflict between supporting traditional leadership and being a committed and active democrat.
Thus, the competition or even contradiction that many assume exists, at least to some degree, between democracy and traditional systems of rule does not appear to exist in the minds of many Africans. Rather than finding themselves trapped between two competing spheres of political authority, Africans appear to have adapted to the hybridisation of their political institutions more seamlessly than many have anticipated or assumed. Chiefs and councillors, sultans and MPs, kings and presidents all inhabit the single, integrated political universe that, for better or worse, shapes each individual's life. In the perceptions of ordinary Africans, it seems that democracy and the chieftaincy can indeed coexist.
OVERVIEW OF THE DATA
This analysis draws initially on data collected between 1999 and 2001 in seven Southern African countries (Botswana, Lesotho, Malawi, Namibia, South Africa, Zambia and Zimbabwe) during Round 1 of the Afrobarometer. The initial conclusions are then confirmed by analysis of additional data collected across fifteen countries in Round 2 (2002–3) (adding Ghana, Kenya, Mali, Mozambique, Nigeria, Senegal, Tanzania and Uganda). Each country is represented by a randomly drawn national probability sample in which each adult citizen had an equal chance of inclusion. Sample sizes ranged from approximately 1,200 up to 3,600 respondents per country, although in the descriptive statistics reported here the data are weighted to represent each country equally.Footnote 1 Women and men were represented equally in each country sample.
Both across and often within the fifteen countries included in this analysis, there can be wide variation in the details of what being a ‘traditional leader, chief or elder’ actually means.Footnote 2 There is variability in what these institutions looked like historically, in what sorts of rules, roles and relationships were imposed on them by both colonial and post-colonial administrations, and in how they have adapted, both individually and collectively, to the many pressures and often competing incentives that they have faced over the years. In the present, there are substantial differences in terms of the extent to which their positions have been integrated into or marginalised from the state bureaucracy, what resources they command, and the nature and extent of both their official and unofficial roles in governing their communities.
But there are commonalities as well. Despite the many ways in which the institutions have evolved over the years, for the most part, people know who we mean when we refer to their ‘traditional leaders’. They generally occupy their posts by virtue of some sort of hereditary (albeit often contested) claim to selection rather than through elections. But more importantly, they are recognised as having connections to their society's cultural and historic roots in ways that official government figures do not. Given the diverse realities that fall under the rubric of ‘traditional leader’, some caution is called for in interpreting the broad cross-country results presented here. But these findings nonetheless offer a valuable first cut at understanding how traditional authorities, diverse as they may be, are faring in the hearts and minds of the average African, especially in an era of advancing electoralism.
TRADITIONAL AUTHORITY IN AN ERA OF DEMOCRATIC CHANGE
Mamdani (Reference Mamdani1996) made the quintessential case against the chiefs, arguing that as long as rural Africans languished under their rule they would remain mere subjects, rather than empowered citizens. Most fundamentally, the absence of voting rights – the most obvious mechanism for ensuring accountability – is seen by many as an insurmountable flaw of institutions of traditional rule. Moreover, in these typically patriarchal systems, the voices of women and youth were largely silenced (Beall Reference Beall2006; Bowen Reference Bowen1994; Molutsi Reference Molutsi and Gyimah-Boadi2004; Simuyu Reference Simuyu, Oyugi, Atieno-Odhiambo, Chege and Gitonga1988). Critics further charge that traditional authority places the community ahead of the individual, is based on a coercive ‘demand for consensus’ and deference, rather than freely given consent, and that, in sum, ‘traditional authority constitutes an anti-democratic, or at best a non-democratic form of governance’ (Mattes Reference Mattes1997: 5–6). From this perspective, these ‘institutional obsoletes’ ‘impede the development of a virile, prosperous, democratic, and just society, and thus must have no place in any progressive society.’Footnote 3 When governments such as the ANC-led administration in South Africa make accommodations with chiefs, critics worry that they are ‘put[ting] at risk the principles of democracy for which [they] fought so hard’ (Beall et al. Reference Beall, Mkhize and Vawda2005: 763–4).
In focusing so intently on the lack of elections, however, detractors often neglect other features of traditional systems that may also be relevant to their democratic compatibility. These include the opportunities they offer for everyday participation (as opposed to periodic voting), as well as their simple familiarity and consequent accessibility (Logan Reference Logan2002). For example, community-wide gatherings common to many African societies, known variously as pitso (Lesotho), kgotla (Botswana), shir (Somalia), baraza (Kenya) and by many other names, have long offered an opportunity for a wide array of community members to voice their opinions on community affairs and participate in consensus-based decision-making. And although many historically excluded women or youth from participation, this norm has been changing rapidly in the post-independence era (Beall et al. Reference Beall, Mkhize and Vawda2005: 770; Chief Linchwe II Reference Chief, Holm and Molutsi1989). Moreover, while heredity typically serves as the basis for selecting leaders, this concept is itself subject to both interpretation and manipulation (Comaroff Reference Comaroff1978), and many systems also had formal mechanisms for ‘de-stooling’ or otherwise displacing leaders who did not meet with the community's approval (Ayittey Reference Ayittey1991: 135–9; Osabu-Kle Reference Osabu-Kle2000: 18). And with their absence of semi-permanent hereditary leaders, the continent's more acephalous systems of rule, for example among the Somali and the Nuer, might be seen as more democratic still (Lewis Reference Lewis1961).
The present reality is, however, far more complex than either of these perspectives allows. After decades of manipulation by colonial and post-colonial governments, and response by indigenous leaders, the content of tradition, and often the identity of traditional leaders themselves, is very often contested (see for example Buur & Kyed Reference Buur and Kyed2006; West & Kloeck-Jenson Reference West and Kloeck-Jenson1999). The machinations of colonial authorities and national leadership have had deep impacts on the status and the very nature of these institutions. Co-optation by colonial governments into the British system of indirect rule, for example, could both strengthen and weaken the hand of traditional leadership, sometimes at the same time. And the efforts of modern African leaders to either undermine traditional leaders and allegiances, or to politicise and thereby capture these potential ‘vote brokers’ (Lawson Reference Lawson2002), have further affected their standing.
The need for chiefs and elders to balance their dependence on local populations for legitimacy, or at least respect, with their desire for the recognition of higher authorities – and the different kind of legitimacy that this implies – may often have produced ‘Janus-faced’ traditional authorities who were simultaneously respected and suspected by local populations (West & Kloeck-Jenson Reference West and Kloeck-Jenson1999: 475–6; see also van Rouveroy van Nieuwaal Reference van Rouveroy van Nieuwaal, van Rouveroy van Nieuwaal and van Dijk1999). Most recently, the chieftaincy in South Africa emerged from under the cloud of apartheid with at best a mixed reputation: some viewed traditional authorities largely as complicit collaborators, while others painted a rosier picture of their essential importance to the stability, solidarity and dignity of their communities.Footnote 4 And while some characterise the manoeuvring of traditional leaders as primarily self-serving, others contest that the ‘ability of chiefs to straddle the state-society dichotomy’ and serve as necessary intermediaries for their people is a strength of the institution that helps to explain its survival (Williams Reference Williams2004: 121; see also van Kessel & Oomen Reference Van Kessel and Oomen1997).
It is hardly surprising, then, that modern African governments have struggled with how best to relate to these institutions. Traditional leaders have both been banned, deposed and jailed, and also courted, coddled and paid state salaries, along with just about everything in between, and sometimes all at the hands of the same governments as these struggle to adapt to their own shifting fortunes. As Lawson (Reference Lawson2002: 9) notes, ‘at the end of the day, the state remained dependent upon traditional authorities for access to rural society’, so it was not uncommon for new and confident administrations to dismiss traditional leadership in their early days, only to come begging for a boost as their popularity sagged in later years.
The question of how traditional authorities ‘fit’ into a modern political system becomes particularly acute at the local level, where these leaders exert the most influence on the daily lives of Africans, and where the contest with government authorities for resources and responsibilities appears to be most intense. While at the national level traditional leaders are often limited to cultural, ceremonial or (frequently undefined) advisory roles, at the community level they may compete with local government officials for real power – over land, tax revenues or other resources, responsibility for dispensing justice, and influence over community activities and decisions, and even votes. The battle between traditional leaders and elected leaders for political power and legitimacy is often characterised as a zero-sum game: whatever authority a traditional leader wrenches from the state is treated as a loss for ‘official’, democratic state leadership, and vice versa. This debate has been particularly heated in South Africa, where the chieftaincy, a ‘ubiquitous feature of local politics’, has been recognised in the constitution and continues to exercise direct authority in many rural areas (Williams Reference Williams2004: 114–16). Chiefs and local government officials are often perceived to be in direct competition – like ‘two bulls in a kraal’ (Oomen Reference Oomen2000: 14) – in a winner-takes-all battle for the hearts, minds and resources of local communities.
In reality, though, it can be difficult to generalise about the relationship between local government authorities and traditional leadership. Even within a relatively small area of South Africa, Oomen (Reference Oomen2000: 62) describes traditional authority areas that range from ‘veritable nations’ to mere ‘backdrops’. And it is not difficult to find examples of constructive relationships between local governments and chiefs who have recognised the mutual benefit that may accrue to both from successful cooperation (see ibid. on South Africa; Owusu Reference Owusu1996: 340 on Ghana).
It is difficult to deny that traditional leaders have demonstrated remarkable resilience. Their continuing importance in the social and political life of their communities, whether perceived as a positive or a negative, is virtually indisputable.Footnote 5 In many places they still play a major role in managing land tenure, adjudicating local disputes, managing property inheritance, and implementing customary law and resolving conflicts. And they are often perceived as the guardians of their communities' culture, playing an important role in cultural events and rituals.
Indeed, traditional authorities may be doing far more than simply surviving. In a number of countries, democratisation appears to be generating a resurgence or revitalisation of the institution. As elected governments must increasingly respond to demands for services with limited resources, they have in some instances come to rely upon traditional authorities to communicate with and mobilise populations (Murray Reference Murray2004; Oomen Reference Oomen2000: 63; see also Englebert Reference Englebert2002; de Sousa Santos Reference de Sousa Santos2006; West & Kloeck-Jenson Reference West and Kloeck-Jenson1999). And the benefits of political liberalisation have applied to chiefs as well. De Sousa Santos (Reference de Sousa Santos2006: 67) notes a ‘growing activism’ on the part of traditional authorities in Mozambique, and they too have the opportunity to form interest groups, such as CONTRALESA in South Africa,Footnote 6 that can advocate on their behalf.
North (Reference North1990) offers some insight into why structures of traditional authority should prove to be so resilient. He argues that the institutions that shape societies – including both their formal rules, and the informal norms and values that underpin them – are relatively resistant to change, in part because of the high information and transaction costs that result. Or as Geddes (Reference Geddes and Crotty1991: 50) describes it, traditional norms and values may be persistent because of ‘the existence of informal sanctions and rewards that persuaded individuals to continue to behave in culturally approved ways’. Moreover, radical changes in the formal rules or institutions of society – such as the introduction of elections and multiparty politics – will not necessarily be matched by corresponding changes in informal norms and values. Informal institutions can and do evolve and adapt, whether in response to changes in the formal rules (or institutions), or shifts in other incentives. But changes will occur incrementally, and at the margins.
Although he is perhaps overly cautious about the potential adaptability of institutions, North's theory has critical implications for democratisation theory and practice. First, it suggests, as we have seen, that existing institutions will not simply melt away, even in the face of alternatives that are clearly ‘better’ in the eyes of their proponents. Second, societies may not understand and respond to newly introduced institutions – such as elections and political parties – in the ways that a less culturally and historically rooted theory would predict. Third, North contends that if there is a lack of correspondence or consistency between formal institutions and a society's informal norms and values, the result will likely be a ‘disconnect’ that produces weak, illegitimate and ineffectual institutions. In sum, North's thesis offers a direct challenge to the assumption that the institutional forms of liberal democracy are easily transferable and universally valid, and the notion that Africans aspire to democratic systems of rule that will look exactly the same as those in the West (see also Ekeh Reference Ekeh1975; Englebert Reference Englebert2000; Landell-Mills Reference Landell-Mills1992).
But the perspectives of ordinary Africans matter more than those of elites or academics on institutions of traditional rule and their place in contemporary African politics. We are just beginning to gather the kind of information that can tell us where chiefs fit in the socio-political constellation of their publics (see for example Beall Reference Beall2006; Williams Reference Williams2004). I now turn to looking at what the Afrobarometer data can contribute to this debate. Are women itching to escape the chauvinistic influence of community elders? Is support for chiefs inherently anti-democratic? Does their continuing legitimacy threaten the consolidation of democracy? Does their exertion of authority undermine democratically elected government officials, especially at the local level?
RATING TRADITIONAL LEADERS: A MODESTLY POSITIVE REVIEW
The seven surveys conducted in Southern Africa as part of Round 1 (1999–2001) of the Afrobarometer provide our most extensive data on popular attitudes towards traditional leaders in Botswana, Lesotho, Malawi, Namibia, South Africa, Zambia and Zimbabwe. It is important to note that this data is also limited in one critical way. Most questions about perceptions of leadership institutions were asked of all respondents. However, the questions about traditional leaders were only asked of those respondents who first answered a filter question about whether they ‘have a traditional leader, chief or headman’ in the affirmative – about two thirds of all respondents in these seven countries.Footnote 7 This ranged from a low of just 19% (408 cases) in South Africa,Footnote 8 to a high of 99% in neighbouring Lesotho (Table 1).Footnote 9
Table 1 Have a traditional leader sub-sample
Question: Do you have a traditional leader, chief or headman?
For those who do have a traditional leader, the survey then went on to ask about that leader's interest in his constituents, his or her trustworthiness, and to what extent chiefs or traditional leaders generally are involved in corruption. Respondents were then asked similar questions about their national president, members of parliament and local government councillors. All of the results presented in this section (on Round 1) only include the subset of respondents who have a traditional leader. Not surprisingly, this filter introduces some bias into our sample, as those who live under a traditional leader are disproportionately rural, older and less educated.Footnote 10 But the analysis of Round 2 results that follows includes all respondents, and it largely confirms the Round 1 findings. So, with this caveat in mind, let us now turn to the substantive findings, starting with the reported perceptions of traditional leaders.
The Afrobarometer began by asking respondents how attentive traditional leaders are to their problems: ‘How interested do you think your local chief is in what happens to you or hearing what people like you think?’ An average of 63% say traditional leaders are either ‘interested’ or ‘very interested’ in their concerns, with one in three (33%) saying they are ‘very interested’ (Figure 1). Namibians show the greatest confidence in the good intentions of their traditional leaders; fully 78% credit them with a real concern for the people. Malawians are not far behind. South Africans, on the other hand, occupy the opposite end of the spectrum: just 48% (of the already relatively small sub-sample) think their own traditional leaders are interested in their constituents' needs and concerns. The question that followed asked respondents: ‘How much of the time can you trust your local chief to do what is right?’ Southern Africans who live under these leaders have moderate levels of trust in their chiefs: 55% say that they trust them ‘most of the time’ or ‘always’ to do the right thing, while just 12% say they never trust them. Again, South Africans are least trusting (35%), and Namibians (72%) and Malawians (61%) are most.
Figure 1 Perceptions of traditional leaders, Southern Africa, Round 1.
Finally, the Afrobarometer asked ‘How many chiefs or traditional leaders are involved in corruption?’ Fifty per cent of respondents believe that the ranks of traditional leadership are relatively free of corruption, responding either that only ‘some, a few’ or ‘almost none/none’ are involved in these illegal activities. Less than one in four (23%), on the other hand, believes that all or most of them are corrupt, while another quarter (25%) say they haven't heard enough about them to say. In Botswana, where traditional chiefs are most integrated into the ‘modern’ political system, we also see the lowest levels of confidence expressed in their integrity. Just 37% think they are relatively free of corruption; however, a surprising 46% say they haven't heard enough about them to say, while just 14% think that many are corrupt. Does this lack of knowledge or an opinion on the matter suggest that Batswana are not as closely linked to their traditional leaders as many analysts have suggested (see, for example, Chief Linchwe II Reference Chief, Holm and Molutsi1989)? Or have Botswana's chiefs, by being integrated into the political system to an unprecedented degree, also been tainted by that system in ways that ordinary Batswana are uncomfortable admitting (see for example Logan Reference Logan2002)?
Factor analysis indicates that responses to these three questions about the quality and integrity of traditional leaders all draw on the same dimension of attitudes or perceptions about traditional leaders.Footnote 11 The three can therefore be combined into a single, five-point Perceptions Index for Traditional Leaders that ranges from 1, equivalent to very negative perceptions of traditional leaders, to 5 for very positive perceptions, with 3 reflecting a neutral attitude.
As indicated in Table 2, the mean value of this index across all seven Southern African countries is 3·4; those who live under traditional leaders thus give them a positive rating overall. South Africans give them the lowest ratings in the region, with a mean of 3·0; in other words, on average, those South Africans who live under traditional leaders (just 19% of all South Africans) sit on the fence in their attitudes towards them, evenly balanced between negative and positive views. In contrast, Namibians, some two thirds of whom live under traditional authorities, give these leaders a strong positive rating with a mean score of 3·9.
Table 2 Perceptions index, traditional leaders
Note: The index averages scores on interest, trustworthiness and involvement in corruption.
THE SELECTED VERSUS THE ELECTED
How do these moderately favourable perceptions of traditional leaders among those who live under their authority compare with assessments of Southern Africa's democratically elected government representatives? Based on the same three questions about corruption, levels of trust and interest in the people, perception indices were created for local government (except for Malawi, and parts of Namibia, which did not have local government bodies at the time of the surveys), members of parliament, and the president and the executive branch.Footnote 12 The results are shown in Figure 2.
Figure 2 Perceptions indices, all leaders.Footnote 19
Comparing the mean values of these indices reveals that, across seven countries, respondents give traditional leaders moderately but consistently better ratings (seven-country mean of 3·4) than any of their elected officials: local government councillors (mean of 3·0), members of parliament (mean of 3·0), and the president and executive branch (mean of 3·1). While the differences are quite small in Botswana and Lesotho, they are much larger in Malawi, Zambia and Zimbabwe; in the latter, a full point separates traditional leaders from parliamentarians and the president. Only in Namibia does any branch – the executive – score higher than traditional leaders, and even then it is only by one tenth of a point.
The minimal differences between traditional leaders and elected officials in both Botswana and Lesotho may reflect the extent of integration of the traditional leadership into the ‘modern’ political system in both of these countries (Englebert Reference Englebert2002: 346), deliberately in Botswana, perhaps more by default in the case of Lesotho. Both are ethnically relatively homogenous societies, and have not experienced the degree of competition between national and local allegiances that may have undermined traditional leadership to some degree in more diverse countries such as South Africa and Zambia. It appears that, in such environments, ordinary citizens may make less of a mental distinction between the government and the traditional leadership.
It is worth nothing that, while South Africans were most critical of their traditional leaders among the seven countries, it turns out that they are even more critical of their elected leadership. Conversely, Namibians are generous not only in their assessments of traditional leaders, but in their evaluation of elected leaders as well. So it is possible that the sizeable difference between these two countries with respect to evaluations of traditional leadership may not reflect vastly different perspectives on traditional leaders specifically, but rather more (South Africa) or less (Namibia) critical inclinations towards leadership in general. I will return to this issue in the following section.
EXPLAINING PERCEPTIONS OF TRADITIONAL LEADERS
Several hypotheses may explain overall perceptions of traditional leaders. The most obvious is the supposition that an individual's level of modernisation will be a good predictor. We might expect younger, more educated, and urbanised individuals to be more likely to endorse the vision of traditional leaders as representatives of a bygone, backward and pre-democratic era, and hence to evaluate them more negatively. Furthermore, many analysts take it for granted that the male-dominated institution of traditional leadership is detrimental to the interests of women (see for example Beall Reference Beall2006); hence, we might also expect men to register higher levels of support for traditional leaders than women. But beyond these initial hypotheses, what else might explain these leaders' ratings?
Using multiple regression analysis, we can both further test the modernisation hypothesis and explore other possible explanations. Additional categories of explanatory factors that have been tested include:
• Performance evaluations: Do popular evaluations of the democratically elected government, its leadership, and its performance on key issues, affect assessments of the relevance and reliability of traditional leaders? If, for example, a person rates elected leaders highly and/or sees the elected government as doing a good job in managing the economy and handling land access issues – traditionally a central realm of chiefly influence – is he or she more likely to see the traditional leadership system as irrelevant, and rate it negatively?
• Trust: Do individuals who express higher levels of general trust in their compatriots also tend to trust traditional leaders more?
• Democratic and electoral attitudes: Are those who support democracy and believe in the principle of elected leadership more likely to reject hereditary chieftaincy? And conversely, are those who are willing to accept non-democratic forms of government more favourable towards what some regard as the authoritarian institution of chieftaincy? Do people who believe that voting can really make a difference in the quality of leadership, or who engage directly in electoral campaigns or rallies, have lower regard for non-elected traditional leaders?
I also test for fixed country effects, which indicate the extent to which characteristics specific to each country (e.g. its own particular history, set of institutions, economic characteristics, or other factors not controlled for in the analysis) are important determinants of perceptions of traditional leaders.
Table 3 shows the results of multivariate regression analysis, with the perceptions index for traditional leaders as the dependent variable. The model explains about 14% of the overall variance in perceptions of traditional leaders, so it offers at least a reasonable start in explaining these perceptions, although there is much that remains unexplained. The modernisation hypothesis does not hold up particularly well, making a negligible contribution to the overall model (adjusted block R square=0·008). Neither age nor urban–rural habitation produce significant effects. Education, in contrast, has significant effects, and they are in the (negative) direction expected, but their magnitude is relatively small. Note, however, that the bias introduced by respondents' self-selection into the study group (see especially endnote 10) could be a significant factor here. Younger or more urbanised individuals who are more negatively disposed towards traditional leaders may simply have selected themselves out of the study group by rejecting the notion that they have a traditional leader. Nonetheless, these findings make clear that, even if such modernisation effects exist, they are far from universal, since the sub-sample analysed here still included sizeable proportions of urban (18%) and younger respondents (69% under age 45), and these appear to be on the same page as their rural and older counterparts.
Table 3 Explaining the perceptions index for traditional leaders, Round 1
* P=<0·05.
** P=<0·01.
*** P=<0·001.
The absence of any significant difference between men's and women's views on traditional leaders is particularly noteworthy, given the obvious concerns about the potentially discriminatory aspects of patriarchal traditional institutions. This could be because, as Beall (Reference Beall2006: 468) suggests, women simply tolerate discrimination because it is part of their culture and tradition. But alternatively (or in addition), it may suggest that we have an insufficient understanding of how traditional systems affect women, and the full range of ways in which women engage with and possibly benefit from these systems – even while facing discrimination in some respects. Ensminger (Reference Ensminger, Nye and Droback1997), for example, has noted that efforts to reform traditional systems of land tenure in Kenya met with community resistance in part because protections for women embedded in the traditional systems were overlooked and eliminated by the reforms. We should perhaps not be too quick to assume that the effects of traditional authority on women are unequivocally negative compared with electoral politics. This is a topic that needs further investigation and analysis.
By far the most powerful set of explanatory factors is evaluation of government performance (adjusted block R square=0·108). But perceptions of government handling of the economy and access to land are not significant. Rather, it is the three other perception indices – for the executive, parliament, and local leaders – that offer considerable explanatory power. But note the direction of these effects: perceptions of traditional leaders are positively linked to perceptions of all other leaders. Moreover, the relationship is strongest with respect to what many describe as traditional leaders' closest ‘competition’: local government officials. In fact, the index of perceptions of local government officials is the single most powerful explanatory variable in our model. In other words, in the battle for the public's favour, not only are traditional leaders not in competition with elected government at the national and especially the local level – recall the image of ‘two bulls in a kraal’ – but, in fact, the two appear to be mutually reinforcing.
There are several possible explanations for this finding. One is that, as suggested above with respect to Namibia and South Africa, individuals may display a leadership disposition or ‘leadership affect’ that is more or less positive, more or less critical, towards all types of leaders – and potentially towards other ordinary citizens as well. Alternatively, whether their role is explicitly recognised or not, this finding may suggest that the public essentially regards traditional leaders as part of a common governing apparatus with elected government authorities, and hence evaluates them on the same criteria as elected leaders. In other words, rather than juxtaposing the government against the traditional system, they may operate based on mental images of a hybrid government. If they are satisfied with the quality of governance, they may attribute this to traditional leaders and government officials equally as components of a single governing system.
The first of these explanations is further supported by the effects of generalised trust, i.e. the feeling that ‘most people can be trusted’. The positive sign on this variable is consistent with the suggestion that evaluations of leadership are based at least in part on an individual's general disposition towards leadership, and towards society at large.
But the second explanation also finds support in the lack of evident incompatibility between public attitudes towards democracy and allegiance to traditional institutions. For the most part, an individual's democratic attitudes – whether pro-democracy, or pro- (or at least tolerant of) authoritarianism – are quite poor predictors of leadership evaluations (adjusted block R square=0·018). Those who support democracy as the best system of government, and those who reject authoritarian alternatives, are neither more nor less likely to have positive views of traditional leaders. The clash that many perceive between traditional rulers and electoral systems of government is not evident to ordinary Africans. And when it comes to engagement with elections, the effects are the opposite of what many would predict. Those who believe in elections, stating that how one votes actually matters to the quality of governance, are more likely to have positive perceptions of traditional leaders, as are those who have worked for a candidate or party. There is thus no evidence to be found here that adherence to traditional leadership and to democratic attitudes are in opposition to one another. Support for democracy readily coexists with support for traditional leadership.
Finally, note that the second most important category of explanatory factors (adjusted block R square=0·058) is the fixed country effect, particularly for Namibia.Footnote 13 This suggests that attitudes towards traditional leaders are not just attributes of particular individuals or groups, but must also be understood as a reflection of the shared experiences of whole populations. All of the countries except South Africa (and Lesotho, where the effects are not significant) tend to be more positive about traditional leaders than Botswana. This is somewhat surprising given Botswana's reputation for providing the paradigm of well-integrated traditional and modern political systems. It is consistent, however, with findings reported elsewhere that integration of traditional leaders into a political system may enhance the legitimacy of the system as a whole, while undermining the standing of the traditional leaders themselves. In such situations, traditional leaders may suffer both from their greater potential exposure to corruption and rent-seeking opportunities, and from the greater distance (both literal and figurative) that their new governmental roles may place between them and their home communities (Logan Reference Logan2002).
RATING TRADITIONAL LEADERS: ROUND 2
In Round 2 (2002–3), only one question evaluating traditional leaders was asked: that of trust.Footnote 14 But it was asked across fifteen countries, and it was asked of all respondents in each country. It therefore provides a broader basis for evaluating perceptions of traditional leaders, and an opportunity to test the initial model of perceptions developed using the more limited (in scope) Round 1 data. Ratings of trust in traditional leaders compared to other elected leaders, as well as several key government institutions, are shown in Table 4.
Table 4 Trust in leaders and government institutions, Round 2
Question: ‘How much do you trust each of the following, or haven't you heard enough about them to say?’ (% a lot/a very great deal).
* Excludes 60% of cases with no local government body.
Once again, traditional leaders fare relatively well in comparison with elected leaders in a number of countries. They get the highest rating across these four groups in eight of the fifteen countries. On the other hand, they receive the lowest ratings among the four in South Africa, Tanzania and Uganda. The biggest surprise, however, is Namibia: for reasons that are not clear, traditional leaders now rank a distant third to the president, in contrast to the neck-and-neck status of the two in Round 1.
We conducted a second multivariate regression analysis using essentially the same model as that developed for the Round 1 Southern Africa data above, but this time applied to all respondents in all fifteen countries, and using trust in traditional leaders as the dependent variable.Footnote 15Table 5 shows the results of this analysis. Note first that, with a more varied sample that includes all respondents (not only those with traditional leaders), the Round 2 model does a considerably better job of explaining the variance in trust in traditional leaders, with an Adjusted R2=0·268.
Table 5 Explaining trust in traditional leaders, Round 2
* P=<0·05.
** P=<0·01.
*** P=<0·001.
Overall, the findings strongly confirm our initial model. The signs and significance of the key explanatory variables, and the relative importance of the performance, democratic attitudes and country blocks are quite consistent between the two models. The most notable difference is that socio-demographic factors have stronger overall explanatory power in the Round 2 model with all respondents included, offering somewhat more support for the modernisation hypothesis than provided by Round 1. Other things being equal, younger, urbanised and more educated Africans are less likely to trust traditional leaders than their older, rural and less educated counterparts, although these factors are still much less powerful in explaining attitudes than performance evaluations. Moreover, this model again shows that gender has no impact on attitudes towards traditional leaders.
The strongest explanations again come from the performance evaluations of other leaders. And as with the first model, it is the ratings, particularly with respect to trust, of local government councillors that act as the most powerful predictors. This confirms the thesis suggested above, that perceptions of traditional leaders are developed largely in conjunction with those of elected leaders – including especially local leaders – rather than in contrast to them.Footnote 16 The common assumption that almost by definition traditional leadership and democratically elected leadership stand in opposition to one another does not hold up.
This model also confirms the findings of the Round 1 model that there is no contradiction between commitment to democracy and confidence in traditional leaders. The explanatory power of democratic attitudes is virtually nil. Even holding a strong belief that elections are the best way to select leaders does not decrease trust in traditional leaders.
As a group, country effects again provide the second most powerful category of explanation. These are significant in all countries except Zambia. There are few surprises, although in slight contrast to Round 1 most countries are more negative about traditional leaders than Botswana. Only Malawi, Mali and Senegal are more positive, which is consistent with observations reported elsewhere (Logan Reference Logan2008) that identify these three countries as home to the staunchest supporters of traditional leadership. Likewise, Uganda and South Africa are, not surprisingly, strongly negative relative to Botswana.
SELECTION, ELECTION AND HYBRIDISATION
To summarise, the data reveal two key findings. First, Africans do not draw as sharp a distinction between their elected and their selected leaders as most analysts and advocates of liberal democracy would expect. Rather than responding to their political environment as though these two groups are in opposition to one another and they must make an either-or choice between them, our respondents draw a strong positive linkage between the two. They appear to evaluate them in the context of a common orientation towards leadership, or ‘leadership affect’, and treat these dual systems of authority as two parts of a single, integrated whole. Second, we likewise see no evidence of a contradiction between allegiance to traditional rulers and individual commitment to democracy. Contrary to the common expectation, those who offer the most positive evaluations of their traditional leaders are no less committed to democracy than those who reject these leaders. Indeed, supporters may even be slightly more active participants in the democratic system.
North's thesis regarding the tendency of norms and values to change gradually rather than abruptly, and the likelihood that new institutions will be integrated into, rather than replace, existing structures, helps us to make sense of these findings. Electing rather than selecting leaders is perhaps the most central tenet of the liberal democratic model of politics; it is non-negotiable. This is precisely what generates so much antagonism in some quarters towards ‘traditional’ systems of ‘hereditary’ (albeit often contested) rule in which elections have no obvious place. But whatever other democratic features they may boast, competitive elections have simply not been a part of Africa's traditional political heritage. North would therefore predict that changing a community's formal institutions – i.e., introducing competitive elections as a mechanism for selecting leaders – would not necessarily produce corresponding changes in the informal norms of society. In other words, granting the public the right to vote will not automatically generate a popular commitment to elections as the ‘only game in town’.
Additional evidence bears this out. Make no mistake: Africans value elections and their increasingly secure right to vote for their leaders. Across eighteen countries surveyed in Afrobarometer Round 3 (2005–6),Footnote 17 for example, 82% agree that elections are the best method for selecting leaders. But they have yet to canonise elections in their political understandings of ‘the right way to govern’ to the extent that the West has. This is evident when we ask people to define democracy (which receives broad support as ‘the best form of government’). When the Afrobarometer asked respondents: ‘What, if anything, does democracy mean to you?’, only 6% of respondents mention voting or elections as their first response, and only another 6% refer to it as a second or third response. Instead, democracy is much more strongly associated with civil liberties and greater personal and collective freedom (31% of first responses). This relatively weak association between the meaning or content of democracy and elections may be due in part to the fact that Africans are still unconvinced about the efficacy of elections. For example, just 47% report that elections ‘enable voters to remove from office leaders who do not do what the people want’. But this lack of efficacy is precisely what North would predict when there is a disconnect between formal institutions such as elections and the prevailing norms and values of society.
The finding that Africans may be somewhat ambivalent – or perhaps, at times, inconsistent – in their attitudes towards elections is confirmed by several local-level studies that have examined the juncture between electoral politics and community norms. In Uganda, Karlström (Reference Karlström1996: 497) reports that in comparing attitudes towards traditional leaders and elected local councillors, the distinction between election and selection ‘was rarely if ever used by my informants in discussing these issues’. When pressed, respondents acknowledged the difference, and even expressed ‘an unqualified preference for popular elections’. But they nonetheless saw the chieftaincy and the council system as more similar than different.Footnote 18 In his study of how elections have been ‘introduced, interpreted and incorporated’ into rural areas in South Africa, Williams (Reference Williams2004: 128) likewise finds that although ‘elections are perceived as a legitimate and necessary mechanism to transfer and distribute political power … it would be a mistake to assume … that people consider elections to be “the only game in town”. Rather, many understand elections to be simply “another game”.’ As a result, Williams' informants weren't just tolerant of institutional coexistence: they expected it.
These processes of institutional evolution and hybridisation are, of course, ongoing – and change is occurring in both directions. Karlström (Reference Karlström1996: 498) notes, for example, that allegiance to elections is growing as Ugandans gain experience with voting as a viable means of achieving society's ‘predominant political ideals’. And Williams (Reference Williams2004: 123) observes that, in the face of pressure from local populations, government, and development agencies, ‘chiefs have sought to “reinvent” themselves both as crucial intermediaries between state and society during formal elections, and as local-level “democrats” who are willing to adapt local institutions in the face of popular pressure for change’. These adaptations have ranged from opening space for greater participation by previously marginalised groups, especially women and younger men, to introducing elections for local chiefs (inzinduna) in some areas. The chieftaincy and democracy are thus engaged in a ‘mutually transformative process [that] illustrates the complexity of democratic consolidation, as well as the ability of the chieftaincy to adapt to changing political and social environments without sacrificing its unique claims to authority’ (ibid.: 113). Similar processes of evolution and transformation have been described elsewhere (Beall et al. Reference Beall, Mkhize and Vawda2005; de Sousa Santos Reference de Sousa Santos2006).
Africans are thus creating for themselves, whether deliberately or by default, hybrid political systems that integrate the traditional systems with which they are deeply familiar and their newly minted electoral regimes. North's analysis suggests that such hybrid systems are an inevitable product of institutional reform. Nonetheless, some fear that any departure from a ‘pure’ model of liberal democracy is likely a prelude to the failure of the democratization project. But the findings presented here suggest that these concerns are overblown. Ake (Reference Ake1991: 34) contends that it is the principles of democracy – including widespread participation, consent of the governed, and public accountability of those in power – that should matter more than the specific set of institutions in which they are embodied. He suggests that we should accept, or even anticipate, that democratic principles ‘may prevail in a wide variety of political arrangements and practices, which naturally vary according to historical conditions’ (see also Englebert Reference Englebert2000; Logan Reference Logan2002; Sklar Reference Sklar and Joseph1999). And Oomen (Reference Oomen2000: 64), too, has observed that if either participation or ‘the ability to debate one's destiny’ can be regarded as an essential principle of democracy, then traditional administrations may often be more democratic than the elected local governments with which they supposedly compete (see also Owusu Reference Owusu1992). Hybridising political systems in these ways may well represent a positive step toward deepening democracy in Africa, rather than a step away from ‘true’ democracy.
All of this highlights the need to further explore and better understand the evolving hybrid political systems that are emerging around the region. A number of analysts have been documenting the complex efforts of various African societies and governments to come to terms with their nations' dual political realities, and explore ways in which the institutions of traditional authority can be effectively blended with the needs of the state, and the principles of democracy. They are also making note of the enormous challenges that these efforts may face (see for example West & Kloeck-Jenson Reference West and Kloeck-Jenson1999 and de Sousa Santos Reference de Sousa Santos2006 on Mozambique; Englebert Reference Englebert2002 on Uganda; Murray Reference Murray2004 and Williams Reference Williams2004 on South Africa). It remains to be seen whether democratic principles will indeed prevail in the various African polities included in this study. But our findings at least indicate that allegiance to traditional rulers is not undermining popular commitment to democracy, and the hybrid political systems that are emerging may be contributing positively to the stability and effectiveness of the democratic political experiment.
* * *
Owusu (Reference Owusu1996: 329) has argued that traditional leadership is such an important part of the local political fabric in Africa that we cannot talk about democracy from below, from the grass roots, without talking about chieftaincy. But there are no simple solutions to the question of how to define the role of chiefs and elders in African political systems circa 2008. De Sousa Santos (Reference de Sousa Santos2006: 66) describes the relations between the parallel worlds of traditional authority and the ‘official politics of recognition and control’ as ‘an intense and chaotic web of interlacings among different legitimacies, local powers, legal cultures, and legal practices’. Separating the two spheres is not an option: ‘this separation, even if correct – which is debatable – is impossible to sustain, given that individuals cannot keep their multiple identities watertight and “uncontaminated”’. Better to recognise ‘that contamination and hybridisation between codes is a “natural” condition’ (ibid.: 61–2, emphasis added). Perhaps, as de Sousa Santos suggests, this integration may lead not to a ‘non-modern alternative to Western modernity,’ but to ‘the expression of a claim to an alternative modernity’ (ibid.: 61, emphasis added).
Liberal democratic purists clearly have legitimate concerns. Traditional systems may be authoritarian and exclusionary. But neither of these is necessarily the case. And focusing too intently on the potential negatives can cause us to lose sight of the fact that traditional systems may also have democratic features, and that they can adopt and are adapting to evolving, increasingly democratic, popular norms. Most importantly, we see that Africans clearly recognise traditional authorities as important actors in their political lives, while also displaying a growing commitment to electoral politics. From their perspective, mixed systems that integrate the two are not just an inevitable but a desirable outcome. We would do well to assume that Africans, especially African women, are not simply deluded in feeling this way, and instead continue to try to understand more fully why this is so. We could begin by exploring further how individuals in Africa define their own vision of a ‘good’ political system, and the ways in which both traditional leaders and electoral politics might contribute towards achieving popular political goals.
Williams (Reference Williams2004: 113, 122) contends that ‘communities seldom believe that they must make an either/or choice concerning democracy and the chieftaincy’, but rather see it as ‘“commonsensical” that the institution of the chieftaincy and democratic elections can, and should, coexist’. The evidence in Africa, as elsewhere in the world, is that societies are often quite adept at integrating seemingly incompatible institutional structures. Citizens in the European and Asian monarchies have been doing it for decades or even centuries, so perhaps we should not be surprised to see that the same may be happening, albeit often on a more local level, in Africa.
