The Study of Term Limits in Sub-Saharan Africa: Lessons on Democratisation and Autocratisation

Abstract

Studies on term limits in Africa have proliferated over the past two decades. This introduction to the special issue on the struggle over term limits contributes to advancing the research agenda with novel empirical evidence and a rigorous conceptual framework. Moreover, we propose complementing existing work on term limits and democratisation with a more explicit focus on their repercussions for authoritarian rule. Drawing on the comparative lessons of the special issue, we outline how term limits can be theorised as part of the institutional landscape in authoritarian regimes and how third-term bids can be understood as a tool of autocratisation.

Introduction

In sub-Saharan Africa, 2020 was another year with a mixed record of term-limit compliance and contestation. Côte d’Ivoire’s and Guinea’s respective presidents Alassane Ouattara and Alpha Condé won controversial third terms. By contrast, Niger’s president Mahamadou Issoufou did not seek re-election for a third term in December. Term limits also featured prominently on the political agenda in Burundi, whose president Pierre Nkurunziza did not run for a fourth term, and in Togo, where Faure Gnassingbé won a fourth term in power. These recent developments show how the struggle over term limits remains a “never-ending debate” on the continent (Mangala, 2020b).

Presidential term limits are constitutional provisions to restrict the number of terms that presidents may serve. They are meant to put a pro tempore constraint on presidential rule and thereby prevent the perils of personal rule and power abuses (Baturo, 2014; Linz, 1998). Within the growing research on term limits in African politics, one strand of studies analyses cross-national patterns of term-limit evasion, manipulation, and abolishment as well as the impact of those tactics on institutionalisation and democratic governance (e.g. Dulani, 2011b; McKie, 2017; Posner and Young, 2007; Reyntjens, 2020; Tull and Simons, 2017). Another strand of research examines more specific aspects of this broader phenomenon, such as protest against third-term bids, international pressure on aspiring overstayers, or legal aspects (see, amongst others, Demarest, 2016; Heyl, 2019; Vandeginste, 2016; Yarwood, 2016) – often by zooming into a specific case. ¹

This special issue further expands these important insights. Rather than focusing on a particular facet of term-limit struggles, its goal is to contribute to some of the key debates that have emerged over the years with rigorous conceptualisations and new empirical insights. We argue that some important distinctions must be made to add further nuance to our understanding of the politics of term limits in sub-Saharan Africa. For one, there is a need to distinguish between two different outcomes that have often been addressed jointly – namely, third-term bids and their success. This also requires disentangling different phases in the struggle over term limits, from the initial decision to contest term limits to the choice of strategy and the success or failure the attempts. These considerations provide the analytical framework for the four empirical articles that make up this special issue. The articles tackle three overarching questions. What are the different ways in which African presidents seek to contest term limits? How successful are they in doing so? And what are the repercussions vis-à-vis democratisation or autocratisation?

The remainder of this introduction first summarises extant research on the politics of term limits in sub-Saharan Africa. It then presents key comparative insights emerging from the four articles in this special issue. A further section discusses how these findings contribute to broader debates about institutionalisation and authoritarian politics. We conclude by sketching implications for the wider literature on term limits beyond the focus on African cases and by outlining avenues for future work in this vibrant field of research.

Existing Research on Term Limits in Africa

Presidential term limits started to catch the attention of scholars studying African politics in the late 1990s when the first African presidents had reached the end of their respective constitutionally prescribed term limits and successfully extended their time in power. Since then, the research on presidential term limits has kept growing in tandem with attempts to violate them. As research and empirical realities developed in parallel, we would first like to situate this research in the historical and empirical context of presidential term limits in sub-Saharan Africa before sketching the state of the art.

Presidential term limits became a widespread constitutional rule in sub-Saharan Africa in the early 1990s, when the third wave of democratisation reached the shores of the continent. However, term limits were not fully foreign to the sub-Saharan African region prior to that. For instance, constitutions in Togo (1963), Senegal (1970), and Mali (1974) already featured presidential term limits (Loada, 2003: 142). Anglophone countries such as Tanzania (1985), Liberia (1986), and Sierra Leone (1985) also introduced term limits prior to the 1990s. However, the pre-1990 term limits did not change the de facto length of presidential rule. On average, African presidents ruled twice as long as their counterparts in other regions of the world between independence and 1990 (Bienen and Van de Walle, 1991). Malawi’s Hastings Banda, Zaire’s Mobutu Sese Seko, and Côte d’Ivoire’s Félix Houphouët-Boigny are examples of such (almost) life-long presidencies.

By the beginning of the 1990s, two-thirds of all African countries with (semi-)presidential systems of government had introduced presidential term limits (McKie, 2017: 439). This development occurred mainly in response to domestic initiatives that sought to create safeguards against the perils of unconstrained personal rule even though term limits did not feature prominently in the democracy promotion agenda of external actors, including the United States Agency for International Development (USAID) (Mangala, 2020b: 6; McKie, 2017: 439). It is not a surprise that term limits were introduced in countries where constitutional reforms were negotiated through inclusive bodies such as national conferences, and where opposition voices had grown strong during the transitional process. Yet, some leaders who possessed a high level of control over the constitution-writing process nevertheless introduced term limits – Burkina Faso, Ghana, Guinea, and Kenya are prominent examples here. In these cases, high levels of electoral uncertainty made the introduction of term limits more likely (McKie, 2017). In some countries, such as Burkina Faso and Senegal, the ruling parties mitigated the constraint posed by term limits by extending the presidential term from five to seven years (Heyl, 2019; Loada, 2003: 143). Between 1990 and 2018, the end of a term limit was reached forty-two times in twenty-eight countries. In twenty-one of these instances, the respective incumbent sought to extend his time in office and sixteen of these attempts were successful (Carbone and Pellegata, 2020: 164). Despite the frequent contestations surrounding term limits, they remain a widespread constitutional phenomenon. As of the beginning of 2020, thirty-five out of thirty-nine African countries with a (semi-)presidential system of government had a term limit in their constitution (Heyl and Llanos, 2020).

Institutionalisation of Power or Continued Personal Rule?

In addition to describing the first instances of term-limit contestations and discussing the (dis-)advantages of term limits, early research on term-limit violations soon began to interpret the meaning of these struggles over term limits in light of the historical legacy of personal rule in Africa (Baker, 2002; Bamfo, 2005; Loada, 2003). The question of whether formal institutions matter in African politics continues to be a major theme in work on presidential term limits in sub-Saharan Africa. It draws on a research tradition that has depicted personal rule by “Big Men” – which is not constrained by formal institutions but upheld by informal practices such as clientelism – as an essential feature of African politics (Chabal and Daloz, 1999; Jackson and Rosberg, 1982). In such a reading, formal institutions – if present at all – are merely window dressing. The empirical realities of the dominant mode in the transfer of power before 1990 support this argument. In the post-independence decades, political power has most frequently been acquired through coups d’état. The majority of incumbents have either died in office of natural causes or were ousted by coups, whereas turnover by way of elections has been a rare exception (Carbone and Pellegata, 2020: 104).

Since the early 1990s, however, formal institutions have gained importance in constraining executives, as Posner and Young (2007) prominently argue. Taking the example of presidential term limits, they show that term-limit regulations were contested in approximately half of the instances in which the incumbent reached the end of the constitutionally prescribed mandate between 1990 and 2005. These challenges were often successful, but term-limit regulations were altered through formal channels such as parliaments or referenda. Thus, rulers did not simply ignore term limits but followed constitutional rules to change, circumvent, or abolish them in order to remain in office, which Posner and Young (2007) interpret as evidence that formal institutions shape the strategies and behaviour of African presidents. In a more recent piece (Posner and Young, 2018), they update and uphold their argument for the period between 1990 and 2018. Their argument speaks to a larger body of literature that posits the increasing importance of formal institutions such as elections and courts (Cheeseman, 2018; Lindberg, 2006; VonDoepp, 2009). Empirical evidence on the dominant mode in the transfer of power also suggests shifts in the relation between formal and informal institutions. Since 1990, almost half of alternations were achieved through elections, while the share of violent takeovers decreased to 18 per cent of all changes in power (Carbone and Pellegata, 2020: 104).

However, the assessment that term limits exemplify the (increased) institutionalisation of power in Africa is not uncontested. Omotola (2011: 124), for example, argues that term-limit contestations constitute a “new form of neopatrimonial rule” and that the use of institutional channels to violate term limits merely reflects neopatrimonialism’s quest for a certain amount of institutional legitimacy. In a similar vein, Tull and Simons (2017) question the strength of formal institutions in sub-Saharan Africa. While their empirical results resemble those of Posner and Young, they interpret the high number of successful third-term bids as a sign that African presidents still enjoy great leverage when dealing with formal institutions. Likewise, Reyntjens (2020: 275) finds that “effective term limits are less prevalent in practice than in constitutional texts.” Whether seeing the glass as half full or half empty, all authors highlight the different trajectories of African countries: in one group of states, term limits are respected, and in the other, term limits are contested.

High levels of institutionalisation and strong formal institutions have long been associated with democracy (Huntington, 1968; Levitsky and Murillo, 2009). Accordingly, many of the studies cited above (Posner and Young, 2007; Reyntjens, 2020; Tull and Simons, 2017) examine how respect for term limits is related to the level of democracy in the respective country. Relatedly, the violation of term limits is perceived as a step in a process of autocratisation (Bermeo, 2016; Omotola, 2011). Nonetheless, a more explicit theorising of term limits with regard to the institutionalisation of authoritarian rule is still in its infancy (for a notable exception, see Meng, 2020), as we further discuss below.

Understanding Term-Limit Contestations and Their Success

With the growth of empirical instances of term-limit contestations, the stocktaking of these attempts has also become increasingly systematic, most notably with regard to the determinants of (successful) third-term bids. This literature identifies the institutional context, the incumbent’s individual characteristics, repression strategies, and societal responses as key factors influencing term-limit contestations and their outcomes. As a baseline, we know that presidents who can count on a cohesive ruling party and a legislative majority are more likely to succeed in extending their time in office (Dulani, 2011b; VonDoepp, 2005). Electoral uncertainty is also an important predictor for the outcome of third-term bids (McKie, 2019). Incumbents are less likely to successfully contest term limits when the level of electoral competition is high. Moreover, strong institutions of horizontal accountability reduce the probability that presidents win third-term bids (McKie, 2019).

Apart from institutional factors, incumbents’ individual characteristics matter. Presidents considered to be father figures – as, for instance, Sam Nujoma in Namibia – are more likely to contest the term limit; older incumbents tend to respect the term limit (Dulani, 2011b; Posner and Young, 2018). Moreover, the president’s military background influences term-limit contestations, as rulers who have served in the military have better access to agents of repression (Dulani, 2011b). In that context, who perpetuates repression is important. While violence committed by party militias does not constitute a promising strategy to secure third terms, state repression can be an effective way to safeguard term-limit violations (Dulani, 2011b). Ethnically stacked armies are more loyal to the president and hence more likely to support third-term bids (Harkness, 2017). In a nutshell, the incumbents’ toolkit for (successful) term-limit contestations should include “money, violence, intimidation, control of state media, and ethnic manipulation” (Mangala, 2020b: 231).

Finally, several studies underline the importance of societal responses to term-limit contestations. This research specifically scrutinises how protest movements form to oppose attempted term-limit violations (Demarest, 2016; Wienkoop, 2020b; Yarwood, 2016). Comparative studies find that countries with a civil society that enjoys a greater degree of autonomy from the government and is able to craft large coalitions with political parties and other societal actors are less likely to experience term-limit violations (Dulani, 2011a; Mangala, 2020a: 231). Afrobarometer data show that the majority of African citizens support term limits (Dulani, 2015), and the level of support is particularly high in countries where the term limit has been violated in the past (McKie, 2020). Leaders who attempted to scrap term limits are thus regularly confronted with strong public dissent, and referenda held to remove term limits have evolved into “focal moments” for mass protest (Carter, 2016: 38).

Enriching Term-Limit Research

This special issue emanates from a double panel at the 2018 conference of the African Studies Association in Germany (VAD) in Leipzig that addressed the struggle over presidential term limits in Africa in an intraregional and diachronic way. The contributions to this special issue build on the Leipzig discussions and approach the study of term limits from a variety of theoretical and conceptual perspectives. Andrea Cassani (2021) conceptualises term-limit manipulations as a mode of autocratisation that allows African presidents to weaken executive constraints and political competition. Based on a three-phase model of what he refers to as term-limit manipulations, he systematically revisits existing explanations of their occurrence and outcome with new data. Anja Osei, Hervé Akinocho, and Stephen Mwombela (Osei et al., 2021) apply Geddes’ typology of authoritarian regimes to argue that presidents in party-based authoritarian regimes are more likely to respect term limits than their counterparts in personalist authoritarian regimes. They illustrate the underlying mechanisms with two case studies, respectively on Tanzania and Togo, and highlight how the well institutionalised ruling party in Tanzania routinely manages leadership transitions, whereas term limits constitute an Achilles heel for the personalist Togolese regime. Daniel Nowack (2021) builds upon the logics of conditionality and appropriateness to analyse international interventions in term-limit struggles. Using the case of Malawi between 1999 and 2003, he shows that bilateral donors and international organisations effectively exerted influence that fostered civil society activism in defence of term limits and led to the erosion of party support for the president’s third-term bid. Finally, Dawson and Young (2021) zoom into the content of constitutional provisions to develop a typology of constitutional term-limit protections. Examining the implications of varying degrees of amendment protection in the cases of Burundi, Rwanda, the Congo-Brazzaville, and the Democratic Republic of the Congo, they conclude that constitutional provisions constrain leaders but not to the expected degree in all four countries.

The special issue also brings together different methodological approaches. Cassani (2021) draws on the African Executive Term Limits dataset and examines the different phases of term-limit contestation by means of a Heckman two-step probit model. Nowack (2021) uses process tracing based on field research and numerous semi-structured interviews. The articles by Osei et al. (2021), and Dawson and Young (2021), occupy a middle ground between the large-N approach of Cassani and the in-depth case study by Nowack. Osei et al. (2021) contrast the cases of Togo and Tanzania, while Dawson and Young (2021) examine four of the most recent cases of Central African presidents facing the end of their constitutional mandates from a comparative perspective.

Despite these different conceptual and methodological approaches, the articles speak to three common themes that are at the centre of term-limit research. First, they examine the different stages of circumventing term limits. In doing so, they build on earlier work by Dulani (2011b: 32–33), who identifies and analytically distinguishes a discussion phase, the tabling of legislation on the removal of term limits, and what he calls the “final outcome.” Cassani (2021) proposes a three-phase model of the manipulation of executive term limits (ETL), consisting of “(1) deciding whether to abide by or try to challenge ETL; (2) choosing how to manipulate ETL; and (3) implementing the strategy.” Similarly, Nowack (2021) examines “the decision to pursue a third term or step down; the choice of strategy when pursuing a third term; and the success or failure of a third-term bid.” In his analysis of what one could describe as the second stage of term-limit manipulations – namely, the choice of strategies – Cassani (2021) applies the distinction between hard and soft contravention introduced by Gideon (2007) and by Posner and Young (2007). Dawson and Young (2021) further differentiate the second stage to disentangle how different degrees of amendment protection constrain the strategic choices of presidents. They argue that constitutional provisions that require the combination of a referendum and a parliamentary vote to change existing term-limit regulations – or what they call “amendment plus” – pose an intermediate level of difficulty for those seeking to extend their terms in office. Such an approach contrasts with a “simple amendment” whereby constitutional term limits can be amended by the same process as any other amendment, and the so-called “amendment block,” which prohibits any changes.

A second way in which the articles of the special issue contribute to the literature on term limits in sub-Saharan Africa is by more systematically distinguishing between different outcomes or dependent variables. The differentiation of several phases of term-limit contestations allows the authors to add nuance to their conceptualisations and empirical analyses of related but disparate outcomes. The articles by Nowack (2021) and Osei et al. (2021) examine both attempts to circumvent term limits and the success or failure thereof. While the authors are thus interested in two different dependent variables, they carefully differentiate them in their analysis. Cassani (2021) goes one step further by also delineating two understandings of “successful” term-limit contraventions:

When we examine success and failure in ETL manipulation, we could assume two different approaches. According to a strictly procedural approach, ETL manipulation is successful every time a term-limited incumbent is able to run again. According to the more substantive approach this paper adopts, ETL manipulation succeeds only when a leader actually manages to hold on to power.

Finally, the contributions to this special issue – like much of the preceding literature – seek to explain the different outcomes of third-term bids. In doing so, the authors both confirm and challenge some of the existing explanations for successful third-term bids, while also discussing additional determinants. Both Osei et al. (2021) and Nowack (2021) confirm that ruling-party cohesion influences the success of third-term bids, as previously shown by VonDoepp (2005). Moreover, Cassani (2021) and Nowack (2021) underline the power of precedent found in prior research (Posner and Young, 2018). The articles also show that certain explanations can be highly context-specific and contingent: Nowack (2021) highlights the role of civil society in fending off third-term bids, whereas Cassani (2021) suggests that the “emphasis previously placed on citizens’ potential as defenders of ETL (Dulani, 2015) should be reconsidered in light of the statistically non-significant results regarding freedom of expression.” Yet this may also be the result of freedom of expression being an imperfect proxy for the civil society pressure identified as crucial in other studies (Demarest, 2016; Dulani, 2011b; Wienkoop, 2020a). In addition to re-evaluating existing explanations with new data and case evidence, the articles also examine explanations that have not received systematic attention thus far, most notably authoritarian regime type (Osei et al., 2021), opposition-party cohesion as opposed to ruling-party cohesion (Cassani, 2021), and the content of the amendment provisions (Dawson and Young, 2021).

Executive Constraints and Authoritarian Rule

In extant studies on the politics of term limits, the level of democracy has been conceptualised both as a factor explaining (the success of) third-term bids (Cheeseman, 2015; Dulani, 2011b; Gideon, 2007; Tull and Simons, 2017) and as the outcome of interest (Baturo, 2014: chapter 8). Reyntjens (2016) problematises this and shows that, on the one hand, democracy is analysed as a predictor of term-limit maintenance, but, on the other hand, there might also be reciprocal causality. Moreover, it is noteworthy that most studies to date have theorised the relationship between term limits and democratic governance even though a majority of third-term bids do not occur in democratic regimes. This close link between democracy and term-limit research can be traced back to the origin of term-limit provisions in sub-Saharan Africa – their introduction during the third wave of democratisation.

Three articles in this special issue move away from this underlying focus on democratisation, each in a different way. The discussion by Dawson and Young (2021) of the four most recent Central African cases shows that many, if not most, of the recent term-limit contestations occur in authoritarian regimes. Osei et al. (2021) build on this observation and argue that typologies of authoritarian regimes provide “more explanatory leverage” for the analysis of third-term bids and their success than does the level of democracy in the respective country. Cassani (2021) conceptualises executive term-limit manipulations as a tool of autocratisation. Taken together, the different articles of the special issue contribute to enriching the research agenda on term limits by theorising and analysing its impact not only on (the lack of) democratic governance but also on different modes and types of authoritarian rule more specifically.

In doing so, they hint at two potential ways in which authoritarianism and term-limit research can cross-fertilise. On the one hand, term limits could be theorised more explicitly as an institution that is not per se incompatible with authoritarian rule (Ezrow and Baturo, 2019; Meng, 2020), thereby contributing to the “institutional turn in comparative authoritarianism” (Pepinsky, 2014). Svolik (2012: 198) argues that term limits allow rulers to signal commitment to constraints on the power of a single ruler. Along these lines, Osei et al. (2021), in this issue, study the relation between ruling-party institutionalisation and term limits. On the other hand, term-limit contestations could be understood as “authoritarian practices” (Glasius, 2018) even if achieved through seemingly law-abiding procedures such as parliamentary amendments, because they have the potential to undermine the voices of those who challenge power-holders. In that vein, Cassani (2021) argues that term-limit manipulations contribute to autocratisation understood as “a process of regime change towards autocracy that makes the exercise of political power more arbitrary and repressive and that restricts the space for public contestation and political participation” (Cassani and Tomini, 2019: 22).

Conclusion

The special issue brings together a group of scholars studying term-limit contestations in Africa from a variety of conceptual perspectives and with different methodological approaches. Despite this diversity, the contributions are united by the analytical rigour with which they differentiate between phases in the struggle over term limits and distinguish third-term bids from their success. The articles comprise rich empirical analyses that leverage new cross-national data, draw on novel interview material, and shed light on very recent cases or those that have received only limited attention thus far. One key innovation of the special issue is its emphasis on the politics of term limits in authoritarian regimes. By moving away from the (implicit) focus on democratisation that has characterised much of the previous literature, the articles hint at new ways to theorise term-limit contestation, whether as an authoritarian practice or as an institution that does not necessarily undermine authoritarian rule – and, indeed, may even be amenable to it.

Presidential term limits are easy to observe and a good indicator of the significance of formal institutions in Africa. Future research should widen the scope of inquiry in two directions. First, a broader approach to the politics of constitutional reforms would allow us to scrutinise whether term-limit reforms are consistent with other reform measures that enhance autocratisation or whether they are part of a controversial negotiation process that may produce contradictory reform packages. Baturo and Elgie (2018) show that authoritarian regimes tend to introduce bicameral legislatures to mask presidential term-limit reforms, but more systematic insights into the interplay of different aspects of constitutional-reform packages are lacking. Second, an analysis of further formal and informal institutional features of presidentialism could expand our knowledge of how executive power is exercised in contemporary sub-Saharan Africa and how this differs from the immediate post-independence decades. Despite the widely recognised continuing central role of African presidents, presidentialism is a largely neglected research field in African politics today.

While the articles of this special issue deal with presidential term-limit struggles in sub-Saharan Africa, their implications for existing and future research are relevant beyond the African continent. Between 1789 and 2006, more than 60 per cent of the constitutions in (semi-)presidential systems worldwide provided for a term limit (Ginsburg et al., 2010: 1836). In countries as diverse as Bolivia, Tajikistan, and China, presidents have successfully contested term limits (Versteeg et al., 2019). And in the United States, where the informal two-term limit was codified only in 1951, Ronald Reagan openly criticised the term limit, and several bills to lift it have been introduced in Congress since, though none were passed (Carbone and Pellegata, 2020: 160; Ginsburg et al., 2010: 1835). Future research can benefit from more cross- and inter-regional comparisons (Ahram et al., 2018), which can challenge premises derived from area-specific patterns in the contestation of term limits.