The (Il)legitimacy of Constitutional Amendments in Africa and Democratic Backsliding

Abstract

In many African countries with hegemonic-party or de facto one-party systems, political leaders have historically exploited ostensibly proper constitutional amendments to undermine constitutionalism, a practice raising questions about the legitimacy, or lack thereof, of such amendments. This article argues that amendment legitimacy is contingent on achieving ‘broad consensus’, a concept endorsed by the African Charter on Democracy, Elections and Governance. Traditional amendment procedures, such as supermajorities and referendums, while crucial, have proven to be imperfect proxies for ensuring such broad consensus. To more effectively safeguard the core constitutional rules of democratic governance, this article contends that political parties must be recognised as key sites of power division and checks and balances. Accordingly, constitutional amendment procedures should require some level of cross-party approval for key amendments, thus preventing individual political groups, regardless of their dominance, from unilaterally altering fundamental rules of the game. This approach would not only enhance the legitimacy of amendments but also serve as a safeguard against contemporary forms of democratic backsliding, where incumbents exploit formal processes to undermine democratic competition. While this process might make constitutional changes more difficult, it would apply only to a narrow set of fundamental aspects of constitutional democracy. Moreover, it does not necessarily conflict with popular self-governance (and its majoritarian expression), but instead calls for an inclusive re-imagining of majoritarianism.

Introduction

The legitimacy of constitutions has become a preeminent intellectual and, increasingly, practical concern in the making and adaptation of constitutions. This concern is exemplified by the United Nations Secretary-General's 2020 Guidance Note on assistance in constitution-making processes, which emphasises the importance of legitimacy of constitution-making and advocates for inclusive, participatory, and transparent processes, which it views as critical to building consensus around constitutional frameworks and ensuring their ultimate success.¹ However, questions about constitutional legitimacy, or lack thereof, are not confined to the initial creation of constitutions, but extend to subsequent amendments or changes to the constitutional framework.

In the African context, the debate over the (il)legitimacy of constitutional change extends well beyond theoretical engagements within academic institutions, and has become a pressing practical issue. While it is methodologically difficult to identify the threshold of public and political opposition that determines the illegitimacy of a constitutional amendment,² legitimacy contestations in Africa have ranged from procedural violations in the drafting of the amendment to conflicts with perceived immutable principles.³ More recently, questions of illegitimacy have frequently arisen in relation to amendments that, at face value, follow established processes, but have been characterised as ‘constitutional coups’.⁴ These instances of ‘constitutional but illegitimate’ amendments force both academics and practitioners to reassess long-held assumptions about the process of constitutional amendments and their legitimacy.

Constitutions, as expressions of the will of ‘We the People’, are seen as embodying a broad social and political consensus, which forms the basis of their (claims to) legitimacy.⁵ Constitutional amendments, upon approval, become part of the constitution and should therefore reflect this broad political consensus. Without such consensus, the legitimacy of these amendments may be called into question.

Traditional constitutional amendment processes typically rely on imperfect proxies to gauge sufficient levels of political consensus and popular support, such as legislative supermajorities (requiring approval from more than half of the total membership) and sometimes referendums.⁶ While these mechanisms may be able to constrain ruling parties in certain contexts,⁷ in practice they often fall short of achieving genuine broad consensus, which, in principle, would require agreement beyond a single political group – especially in many African countries with hegemonic-party or de facto one-party systems. As a result, constitutional amendments in these contexts frequently face contestations over their legitimacy.

After identifying the common sources of legitimacy challenges to constitutional amendments in the African context, this article argues that the legitimacy of constitutional change should be measured by the extent to which such change results from a process aimed at achieving broad consensus. Using broad consensus as the benchmark, the article contends that the current reliance on standard amendment procedures, particularly supermajorities and referendums, has allowed dominant groups to push through capricious, self-serving, and generally regressive/abusive⁸ reforms. A pragmatic approach to achieving broad consensus, and thus legitimacy, may be to supplement these imperfect procedural proxies with cross-party approval requirements. Under this approach, amendments would only be valid if supported by a majority (or a specified proportion) in a majority (or a specified number) of political groups in parliament, rather than merely a supermajority of the total membership – a threshold often easily met by dominant groups in Africa. The core argument is that no single political group, regardless of its electoral dominance, should be able to unilaterally alter the fundamental constitutional rules of the democratic game.

This proposal can be seen an adaptation of the ‘separation of parties, not branches’ idea propounded by Levinson and Pildes.⁹ It considers parties as the primary arenas of political contestation and checks and balances. Given that the proposed amendment process will make constitutional change particularly cumbersome, it must be applied selectively to the most vulnerable constitutional provisions – the so-called democratic ‘minimum core’¹⁰ as determined in each constitution, which will inevitably vary by context. In the African context, this ‘minimum core’ could encompass provisions on presidential term limits and the electoral system, safeguards for judicial independence, appointment procedures for judges and members of other key ‘fourth branch’ accountability institutions (eg, electoral commissions), and protections for democratic rights and minorities. As such, the proposed amendment procedure would represent one end of the spectrum on the ‘entrenchment escalator’.¹¹

In sum, the proposed procedure, which could be termed an ‘inclusive majoritarian amendment process’, seeks to enable legitimate constitutional change by requiring broad – ie, inclusive – consensus, without unduly preventing the adoption of necessary reforms. This approach ensures that no single political group can unilaterally alter fundamental aspects of the constitutional framework, thereby mitigating the risk of democratic backsliding. It also addresses potential objections to the idea of unamendable constitutional provisions and other substantive limits on constitutional change, rooted in the notion of a majority's right to ‘self-governance’. While it does not explicitly affirm the existence of unamendable provisions, it achieves a similar outcome – not by rejecting or restricting majoritarianism, but by calling for a rethinking of what constitutes a majority in order to achieve genuine broad consensus.

It is important to note that the proposal for cross-party approval primarily targets aspects of the constitutional framework that are crucial to the functioning of a democratic system. These aspects have to be identified for each constitution individually, based on its unique history and context, and are not applicable to all constitutions as a whole. While the proposal is certainly not a panacea, it serves as a vital safeguard against illegitimate amendments. Although this article focuses on political parties, the strength and resilience of a democratic framework ultimately hinges on a combination of popular vigilance, a vibrant civil society, and supportive courts. Without these, a functioning constitutional democracy is difficult to imagine. Indeed, strong opposition parties, vigilant citizens, an engaged civil society, and independent courts should be viewed as complementary and mutually reinforcing pillars of democracy, not as alternatives.

Sources of Illegitimacy of Constitutional Amendments in Africa

Illegitimacy of constitutional amendments, like issues of illegitimacy more generally, takes different forms. In the African context, questions about the legitimacy of constitutional amendments have arisen from a range of issues related to process, substance, and context. Claims of illegitimacy have often concerned the exclusion of opposition parties from the amendment process, the use of corruption to secure support for amendments, and, in some cases, the outright use of violence against opponents.

Illegality as Illegitimacy

The first source of illegitimacy of constitutional amendments is the violation of the amendment procedure provided for in the existing constitution. This draws on Max Weber's characterisation of the belief in the law (legality) as the most common form of legitimacy.¹² While constitutional change in Africa is increasingly taking place through prescribed procedures, which has been described as one sign of the ‘institutionalisation of politics in Africa’,¹³ there are recent examples of constitutional change being achieved outside the constitutional framework. For instance, in 2015, the President of the Republic of Congo, Denis Sassou Nguesso, orchestrated the creation of what was ostensibly a new constitution through a referendum in order to bypass the constitutional prohibition on amending or removing presidential term limits.¹⁴ The replacement constitution allows individuals to serve up to three prospective presidential terms (Article 65), not counting terms already served. Following the same script, President Alpha Conde of Guinea-Conakry orchestrated the making of a new constitution in 2020 that allowed him to bypass the provision prohibiting the amendment or removal of presidential term limits.¹⁵ In both countries, the French tradition of pouvoir constituant (constituent power) – the inherent power of the people to make, unmake, and remake the constitutional framework – was conveniently invoked to justify the violation of constitutional legality.¹⁶ It is important to note that the processes for making the replacement constitutions, which were not regulated in the older constitutions, were more executive-dominated, with final adoption by referendum, than the burdensome amendment rules in the older constitutions. Amendments to key constitutional provisions would have required a significant legislative supermajority, and approval by referendum.

In sum, invoking constituent power to bypass the established constitutional framework and pursue a constitution-making process that disregards the basic constraints of constitutional amendment processes offers incumbents – and only incumbents – unbounded discretion, undermines legality and legal certainty, has been prone to frequent abuse, and should therefore be considered illegitimate.

In contrast to successful extra-constitutional reform processes in Congo and Guinea, Niger's President Mamadou Tandja faced a different outcome in 2009. Tandja sought to extend his presidential term by proposing a referendum to adopt a new constitution, but the National Assembly rejected the proposal. In response, Tandja dissolved the Assembly by force. When the Constitutional Court subsequently ruled that the referendum proposal violated the constitution, Tandja responded by banning the Court and proceeding with a rigged referendum regardless, which approved the new constitution. However, within a year, the military overthrew Tandja and transferred power to a civilian transitional government, leading to the adoption a new constitution in 2010, which reinstated term limits.¹⁷

It may be argued that the constitutional legality of an amendment, as the most traditional source of legitimacy, does not necessarily require that the original constitution is legitimate, and that the fact that an amendment is challenged as illegitimate does not necessarily imply the legitimacy of the original constitution. For instance, the transition from Apartheid to a democratic dispensation in South Africa was pursued within the formal framework of the Apartheid constitution.¹⁸ The 1993 Interim Constitution, which incorporated the essence of the agreement between the Apartheid government and the African National Congress, was adopted by the Apartheid parliament and through procedures established in the then constitution. Similarly, in Kenya in 2008, there was consensus that the previous constitution required reform and a new constitution needed to be enacted.¹⁹ Accordingly, a process for the making of a new constitution, which was not expressly regulated by the existing constitution, was first agreed upon by the main political actors.²⁰

Irrespective of the legitimacy of the existing constitutional framework, constitutional legality is often emphasised in ‘pacted’ transitions from authoritarianism or conflict to democracy or peace, such as in South Africa's shift from Apartheid to democracy in the early 1990s and in Kenya after the post-election violence in 2007. In contrast, when transitions result from rebel victories or popular uprisings, the existing constitutional framework may be set aside, and the legitimacy of the new framework is drawn from sources outside the existing constitutional framework. For instance, following the overthrow of Omar al-Bashir in Sudan in April 2019, the constitution was simply discarded, and a transitional constitutional declaration was agreed upon by the military and civilian protest leaders.²¹ Similarly, following the removal of Ben Ali from power in Tunisia during the 2011 popular uprising that kickstarted the Arab Spring, the existing constitution was suspended and a transitional arrangement was put in place, eventually leading to the adoption of a new constitution in 2014.²² In such contexts, legitimacy is seen as independent from legality.

Constitutional but Illegitimate

While illegality can be a useful indicator of illegitimacy, it does not fully define it. Constitutional amendments may be legally and procedurally sound, but their legitimacy may still be questionable. This can occur if the amendment process is seen as inadequate, or if the amendment has been implemented in a way that contradicts the spirit and intent of the constitution's drafters as to the function and purpose of the amendment procedure established in the constitution – for instance, constitutional supermajority requirements presuppose competitive elections that make it difficult for any single group to easily secure the high numbers needed to alter fundamental aspects of elections unilaterally, which is intended to necessitate deliberation and moderation among diverse groups. Beyond these procedural considerations, questions of legitimacy may also arise from the substance of the reforms or their potential consequences.

In the first instance, an amendment has been enacted in accordance with established constitutional procedures, but is still deemed illegitimate due to the surrounding circumstances or context. This is particularly the case when a single political group dominates the political process, making the reform inherently non-inclusive and lacking in broad participation. A notable example is the 2017 Ugandan constitutional amendment, where President Yoweri Museveni orchestrated the removal of the presidential upper age limit, allowing him to run for office again (term limits had already been removed in 2005).²³ After being passed by a two-thirds majority in parliament, the amendment faced widespread criticism for its perceived illegitimacy, as all opposition groups in parliament rejected it. It was seen as a capricious reform pushed through by a single political group. Claims of illegitimacy were further fuelled by public opinion polls showing overwhelming rejection of the amendment.²⁴ Despite calls to submit the amendment to a referendum, the government refused, arguing that this was not constitutionally required. Instead, the government provided funds to members of parliament to supposedly engage with their respective constituencies about the amendment – a move the opposition decried as a corrupt attempt to buy support.²⁵

Increasingly, many African countries have engaged in processes beyond the constitutionally prescribed amendment procedures to enhance the legitimacy and technical quality of the proposed reforms. Notably, expert constitutional commissions have become increasingly common, even though their role is not explicitly anticipated by most constitutions. For instance, in 2013, the Senegalese President established the National Commission for Institutional Reform (Commission nationale de réforme des institutions, CNRI). This commission proposed several constitutional reforms,²⁶ some of which were eventually put to a referendum and approved in 2016, including reducing the presidential term from seven to five years. Similarly, The Gambia established a Constitutional Review Commission in 2018 to guide a participatory and inclusive constitutional reform process.²⁷ Botswana also established an expert commission to consult the public and propose reforms.²⁸ These extra-constitutional procedures aim to ensure expert involvement in constitution-making, garner broad political and popular support, and bestow greater legitimacy on the proposed reforms.

Some constitutions may also enjoy such strong popular legitimacy that any proposed constitutional change faces questions of legitimacy.²⁹ This is evident in several Francophone African countries where ‘Do Not Touch My Constitution’ (Touche pas à ma Constitution) movements have effectively thwarted reform efforts, particularly on constitutional terms, but also more broadly on presidential powers.³⁰ In Mali, for example, this movement has long prevented any amendments to the 1992 Constitution, despite multiple attempts at reform. On at least two occasions, comprehensive reform proposals were overwhelmingly approved by parliament, only to be abandoned shortly before being put to a popular referendum.³¹ The high original legitimacy accorded to these constitutions has created a presumption that any reform is illegitimate, triggering intense popular and political resistance. This resistance has been particularly strong because reform proposals have been seen as pretexts for extending presidential terms, or, in the most recent case, as attempts to constitutionalise the outcomes of the Algiers Peace Agreement between the Malian government and Tuareg rebels.³² Opposition groups and the broader public perceive this peace agreement as the result of exclusive negotiations without input from the public or opposition, and as heavily foreign-led. Attempts to constitutionalise the agreement through amendments are thus seen as legitimising the outcome of an illegitimate peace process.

The legitimacy of constitutional amendments may also be questionable where the president bypasses parliament and submits proposed changes directly to referendum.³³ For instance, in July 2018, President Azali Assoumani of the Comoros relied on a controversial constitutional provision to ban the Constitutional Court, bypass parliament (which would likely have rejected the reforms), and submit the proposed amendments to a referendum.³⁴ The amendments allowed the president to stand for re-election and altered the system of rotating the presidency among the political leaders of the country's three main islands. Assoumani was subsequently re-elected. A similar situation unfolded in Burundi following the disputed re-election of President Pierre Nkurunziza in 2015:³⁵ a commission for national dialogue was established in 2017 and, after ostensible consultations, the commission submitted its report and proposed constitutional reforms to the president. The reforms were approved in a referendum in May 2018, bypassing parliament entirely – even though it was dominated by the president's supporters. This strategy was likely aimed at avoiding a repeat of 2014, when the president's attempt to extend term limits through constitutional amendment was scuttled by a single vote in the Senate.³⁶ Following this setback, Nkurunziza had turned to the Constitutional Court for a ruling that his first term, when he was elected by parliament as a transitional measure rather than through popular elections, would not count towards his term limits.³⁷ Although the revised constitution did not specifically count terms served, the president did not run for another term in May 2020.

Constitutional amendments adopted through established procedures may still be deemed illegitimate if they are perceived to undermine basic (democratic) principles. This situation involves balancing procedural with normative or substantive conceptions of legitimacy, which may sometimes be at odds. Some constitutions explicitly protect certain substantive principles from constitutional amendment, thus creating unamendable provisions.³⁸ These are particularly prevalent in Francophone and Lusophone Africa, but less common in Anglophone Africa.³⁹ For example, while the South African Constitution identifies several fundamental principles and values to guide governance and constitutional interpretation, it does not explicitly preclude their amendment – but it establishes a strict amendment procedure. Debate exists over whether South Africa's Constitutional Court could invoke the ‘basic structure doctrine’ to reject amendments deemed incompatible with unspecified fundamental substantive principles. However, given that the constitution does indeed identify fundamental principles and indicates their amendability, the Court is unlikely to endorse such substantive limits on amendments.⁴⁰

To date, no highest court in Africa has endorsed the idea of ‘unconstitutional constitutional amendment’ to invalidate procedurally valid amendments based on implied substantive unamendable principles, ie, outside the context of express unamendable provisions. Indeed, the Kenyan Court of Appeal had annulled constitutional amendments not only on the grounds of procedural violations, but also on the grounds that they would violate the ‘basic structure’ of the 2010 Constitution.⁴¹ However, the Court did not assert that the basic structure cannot be amended. Instead, it set out a series of steps to be followed beyond the express amendment procedure provided for in the Constitution. Subsequently, the Supreme Court of Kenya upheld the invalidity of the proposed constitutional amendments on procedural grounds, including limitations on the president's power to lead citizens’ initiatives for amendment and failure to ensure adequate public participation in the process, but the Court reversed the lower courts’ endorsement of the ‘basic structure doctrine’ and the four-step process they had proposed.⁴²

The idea of substantive unconstitutionality of constitutional amendments, although rarely recognised and enforced by the courts, finds expression in the African Charter on Democracy, Elections and Governance (ACDEG). This charter prohibits unconstitutional changes of government,⁴³ including ‘any amendment or revision of the constitution’ that constitutes ‘an infringement of the principles of democratic change of government’.⁴⁴ However, this provision has primarily been invoked against coups d’état, and the African Union has yet to employ it to reject constitutional reforms as unconstitutional changes or retentions of government power.⁴⁵ The closest the African Union came to invoking this prohibition was during the Burundian crisis in 2014, when the President sought a Constitutional Court interpretation allowing him to run for a third term, which was seen as a form of judicial amendment of the constitution.⁴⁶ Although the African Union initially rejected the move, it did not frame its objection within the unconstitutional change of government doctrine. Subsequently, the African Union softened its stance, refraining from threatening Burundi with suspension or other punitive measures typically associated with unconstitutional changes of government.

Continental human rights and other African and international standards, it can be argued, also play a crucial role in the assessing the substantive legitimacy of constitutional amendments. In particular, the African Court on Human and Peoples’ Rights has found a Tanzanian constitutional amendment banning independent candidates from contesting presidential and other elections to be incompatible with the right to political participation.⁴⁷ Similarly, the Court has ruled that the composition of the Electoral Commission of Cote d'Ivoire failed to meet requirements of independence and impartiality of election management bodies as guaranteed in the ACDEG.⁴⁸ And in 2020, the Court found a constitutional amendment in Benin invalid on the grounds that it violated the principle of ‘broad consensus’. This was due to the fact that the parliament that approved the amendment was fully controlled by parties affiliated with the President of the Republic, and that there had been no broad consultation.⁴⁹

However, the involvement of the African Court and continental standards in assessing the validity of constitutional amendments, while defensible on grounds of legality and substantive understandings of legitimacy, might be seen as undermining reforms that are domestically perceived as legitimate, including those enacted with cross-party and popular endorsement. This raises questions about the geographical parameters in understanding legitimacy. Notably, in the Tanzanian case discussed above, the Tanzanian Supreme Court upheld the validity of the constitutional amendment on the grounds that the constitution allows for the amendment of any provision as long as the procedural requirements are met.⁵⁰ The Court expressly considered and rejected the ‘basic structures doctrine’, as conceptualised in India, as imposing extra-constitutional constraints on the amendment power.⁵¹ The Court nevertheless acknowledged the potential conflict of the ban on independent candidacy with Tanzania's continental and international human rights obligations and recommended that the relevant political actors reconsider the ban.

The Tanzanian case raises first-order questions about the very existence and origins of extra-constitutional substantive limits on the amendment power, their normative justification, and, critically, which actors have the authority to identify and enforce them. The recognition of the existence of extra-constitutional limits on the amendment power does not necessarily imply their judicial enforceability. Indeed, in many African constitutions, constitutionally recognised principles and provisions are expressly excluded from judicial enforcement, as is often the case with directive principles of state policy guaranteeing socio-economic entitlements.⁵²

While international instruments may provide one source of extra-constitutional limits on the amendment power, they face challenges of legitimacy.⁵³ Crucially, in many countries, international instruments are subordinate to the constitution, even in states with a ‘monist’ tradition where international law can be directly applied once endorsed by the state.⁵⁴ For instance, in many Francophone African countries, which generally follow a monist tradition, constitutional courts must assess the compatibility of international agreements with the constitution prior to their approval. Consequently, international instruments may not provide an adequate basis for determining the validity or legitimacy of constitutional amendments.⁵⁵ Notably, while international law prohibits using domestic law to justify violations of international obligations,⁵⁶ this principle is primarily enforceable at the international level, and domestic courts tend to be reluctant to elevate international obligations over the constitution, unless there is an express constitutional provision to that effect.

The idea of substantive illegitimacy of constitutional amendments is closely tied to the exclusion of incumbents from benefitting from such reforms. Amendments that directly advantage a sitting president or legislature may be opposed as illegitimate, particularly in African contexts where traditional checks and balances are not always effective, and even elections and referendums may be subject to manipulation. Some constitutions have recognised the questionable legitimacy of self-serving amendments and expressly preclude incumbents from benefitting from constitutional reforms made during their time in office. A notable example is the Zimbabwean Constitution, which states that an amendment extending term limits of a public office does not apply to any person who held that office, or an equivalent thereof, prior to the amendment.⁵⁷ This approach offers a helpful way to preclude directly personal amendments, notably in relation to term limits. But its utility is limited because it cannot prevent constitutional change that gives undue advantage to incumbent political groups, such as institutional reforms related to the judiciary or the electoral system.⁵⁸

Broad Consensus as Foundation of Constitutional Amendment Legitimacy in Africa

The discussion and examples above provide the various circumstances in which constitutional amendments have been challenged as illegitimate in the African context. Assessing the justifiability of these claims requires establishing criteria to determine the conditions under which an amendment may be considered illegitimate. This necessitates a thorough examination of the rationale for allowing constitutional amendments and the goals amendment processes seek to achieve, which must be linked to the sources of legitimacy of constitutions themselves.⁵⁹

Modern constitutions often derive their legitimacy and authority from the notion that they constitute expressions of the will of ‘We the People’ of the country, alongside arguments grounded in reason and justice.⁶⁰ In practical terms, the legitimacy of constitutions arises from a broad political and/or popular consensus.⁶¹ In certain cases, this broad endorsement may be achieved at the time of constitution-making, while in others legitimacy emerges or is strengthened over time through the establishment of an expectation of binding compliance. In the African context, countries have increasingly established elaborate procedures for the making of constitutions, including frameworks for inclusively engaging and garnering support from a wide range of political, socio-cultural, and economic elites, as well as ensuring participation of the broader public.⁶² Elite support is often sought through constitution-making processes that include either significant political groups or the establishment of independent expert commissions to engage all relevant stakeholders. Meanwhile, popular buy-in is often sought through increasingly participatory processes and/or by subjecting the final draft constitution to a popular referendum.⁶³

Broad political consensus as a basis for constitution-making is as old as the making of the United States Constitution. While there were no political parties at the time, the draft constitution was unanimously adopted by the representatives of the then thirteen states and came into force after approval by nine state conventions, applying only to those states that approved it.⁶⁴ More recently, the first phase of the negotiation and drafting of South Africa's Interim Constitution, in particular the binding principles that guided decisions on the content and process of the final constitution, was founded on the principle of ‘sufficient consensus’ in a multi-party platform, which, while far from requiring unanimity, necessitated agreement among the major political players.⁶⁵ This commitment to inclusivity was confirmed by the author's personal interviews with individuals directly involved in drafting South Africa's constitution: Nelson Mandela reportedly instructed delegates from the dominant African National Congress (ANC) that they were drafting a constitution for South Africa, not for the ANC. Similarly, the Constitutional Court of Benin has recognised ‘national consensus’ as the foundation of the country's constitution, which was adopted under the auspices of a large ‘National Conference’ representing a broad range of Beninese society.⁶⁶ The Court invoked this principle to invalidate a constitutional amendment adopted in a closed parliamentary session that extended the term of parliament from four to five years,⁶⁷ ruling that the consensus underlying the Constitution required any amendment to be adopted through an open, participatory, and consultative process.

The United Nations guidelines on constitution-making emphasise ‘national ownership’ of the process and stress the importance of inclusivity, participation, and transparency as key determinants of the acceptability and legitimacy of the final constitution.⁶⁸ Similarly, the African Union Peace and Security Council Policy on Post-Conflict Reconstruction and Development (2006) outlines principles for constitution-making, including the pursuit of participatory processes.⁶⁹ However, this policy specifically applies to post-conflict countries, meaning there is no overarching African standard for assessing the legitimacy of new constitutions across all contexts.

There are, however, certain standards regarding constitutional amendments. In particular, the ACDEG requires states parties to ensure that ‘the process of amendment or revision of their constitution reposes on national consensus, obtained if need be, through referendum’.⁷⁰ The African Union Assembly of Heads of States and Governments has also called on states to ensure that ‘constitutional amendments are done in accordance with the provisions of the ACDEG, as a baseline, and the active participation of their citizens’.⁷¹ Like constitution-making, constitutional amendment procedures should not only ensure that various political groups and the populace participate, but also that the outcome represents broad consensus. Beyond Africa, the Venice Commission on Democracy through Law, in a 2022 report on the draft Constitution of Chile, expressed a similar understanding that ‘the adoption of a new and good Constitution should be based on the widest consensus possible within society’.⁷² At a minimum, these standards arguably require an amendment process that ensures reasonable political support beyond the core of the ruling party or coalition⁷³ and allows the public a reasonable opportunity to express their views on the amendments, including through referendums if necessary.

As can be seen, constitutional amendments, much like new constitutions, derive their legitimacy from the extent to which they are the product of a broad national consensus. Constitutional amendment procedures are used as simple and predictable, but imperfect, proxies to ensure broad political and popular support for amendments. The most common requirement for constitutional amendments, both in Africa and elsewhere, is the approval by a legislative supermajority.⁷⁴ In addition to, or as an alternative to, legislative supermajorities, some constitutions require certain constitutional amendments to be approved through a referendum. There are also constitutions that preclude certain provisions from constitutional amendment, or that prohibit amendments during emergencies and similar circumstances.

Supermajority requirements are based on the premise that, given a level of competitive democracy, opposition parties will secure enough legislative seats to be able to block amendments pushed exclusively by the ruling political group.⁷⁵ These requirements are intended to ensure that decisions take into account a range of interests and to promote deliberation, persuasion, and compromise.⁷⁶ By doing so, they offer a liberal, identity-blind mechanism for achieving broad political consensus on constitutional amendments. In cases where the ruling group holds a large majority, opposition views would be considered fringe, and would therefore not taint the existence of a broad consensus. On the other hand, where additional referendums are required, the assumption is that a general popular mandate to govern does not necessarily extend to a mandate for specific constitutional amendments, which are considered to require more direct popular endorsement.

However, the assumptions underlying established proxies for broad political and/or popular consensus often fail to hold in practice, particularly in post-colonial African states. Since the independence period in the 1960s, African incumbents have manipulated legislative and popular majorities to push through constitutional amendments that clearly do not represent broad political and public consensus.⁷⁷ This manipulation occurs alongside frequent violations of constitutional standards and the establishment of parallel informal constitutional arrangements, leading a prominent scholar to decry the prevalence of ‘constitutions without constitutionalism’.⁷⁸ Examples of this phenomenon abound. In 1971, soon after establishing a one-party state, the Malawian legislature declared Hastings Banda, who had led the country since its 1964 independence, ‘President for Life’, exempting him from the need to run for future elections. Banda's thirty-year-rule ended only in 1994, when he lost a presidential election following the reinstatement of multiparty democracy. A few other presidents followed Banda's example and declared themselves presidents for life. In 1982, Kenya's dominant ruling party amended the constitution to declare the country a one-party state and banned competing political groups, further consolidating power.

The ‘Third Wave’ of democratisation in the 1990s (arguably the first in the African context) led to the adoption of constitutions that garnered broad political and popular support, such as in Benin (1991) and South Africa (1996). This trend continued in the 21^st century with constitutions like those of Kenya (2010) and Niger (2010). Nevertheless, the formal return to multiparty democracy did not put an end to the manipulation of existing institutions to push through abusive and capricious constitutional amendments of questionable legitimacy. This is not to discount instances where constitutionally established procedures and the relative legitimacy of these constitutions have successfully blocked self-serving constitutional changes.⁷⁹ Procedural safeguards, the regular political process, and popular opposition have at times successfully thwarted efforts to adopt self-serving constitutional amendments. For instance, attempts to remove presidential term limits in Zambia (2001), Malawi (2002), Nigeria (2006), Benin (2006, 2009), and Burkina Faso (2014) were defeated. Similarly, in a few cases, the people have rejected efforts to change constitutions through referendums, as seen in Seychelles (1992), Zimbabwe (2000), Kenya (2005), and Zambia (2016). These rejections were often due, at least in part, to the perceived illegitimacy of the proposals, which, despite initial popular and political engagement, were eventually hijacked by the ruling parties.

Nevertheless, despite opposition, African incumbents have on many occasions succeeded in pushing through constitutional changes via established procedures. The standard amendment procedures have been unable to secure the broad national consensus they were meant to ensure. Nothing illustrates this as much as the frequent tampering of the length and number of presidential terms. As noted above, Ugandan President Yoweri Museveni, in power since 1986, orchestrated a 2005 constitutional amendment through a referendum to remove term limits. In 2017, he led the removal of upper age limits for presidential candidates to allow him to run again, while reinstating the two-term limit. Similarly, Chad's President Idris Deby, who was in power since 1990, orchestrated the removal of the two-term presidential limit in 2005 and reinstated it in 2018 (he died a few weeks after winning his sixth term in April 2021). Former Algerian President Abdelaziz Bouteflika oversaw the removal of term limits in 2008, only to reinstate them in 2016, both times through parliament approval. Authoritarian leaders in Gambia (2001) and Gabon (2003) oversaw the replacement of the two-round run-off system for presidential elections with a plurality system. These examples illustrate the normalisation of ‘constitutional politics’, where constitutions are amended or replaced with an alarming ease, more akin to ordinary legislative processes, despite supermajority and other requirements.⁸⁰

This history of regressive amendments in Africa reveals that supermajorities and referendum requirements are inadequate proxies for ensuring broad political and popular consensus in constitutional amendments. Similar concerns apply to constituent assemblies, which will inevitably mirror the political factionalism present in regular representative legislative bodies.⁸¹ Scholars and practitioners involved in constitution-making and change should therefore explore mechanisms that can more effectively ensure that constitutional amendments, particularly those affecting core aspects of the democratic process, truly reflect broad national consensus.

Operationalising Broad Consensus: From Supermajorities to Inclusive Majoritarianism

Given the landslide electoral results of dominant incumbent political parties in Africa, often achieved through a combination of abuse of state resources, rigging, and sometimes outright suppression, it is evident that reliance on supermajorities and even referendums has limitations in scuttling abusive and illegitimate constitutional amendments. While amendment rules are crucial, especially in countries with competitive elections and relatively functional parties,⁸² the frequent abuse of even the most rigid amendment procedures supports Ginsburg and Melton's observation that ‘constitutional culture’ is often more influential in determining how frequently constitutional amendments occur than the formal rules governing such amendments, regardless of whether those rules are flexible or rigid.⁸³ This requires a reimagining of standard constitutional amendment procedures, as assumptions about the plurality of legislatures based on competitive elections and credible referendums do not always obtain. This article argues that constitutional designers must pursue more direct ways of ensuring broad political consensus for constitutional amendments, rather than relying on imperfect proxies like supermajorities and referendums.

Adapting the idea of ‘separation of parties, not branches’,⁸⁴ this section proposes an amendment procedure that would require amendments to the core aspects of the constitutional democratic framework to be approved on a cross-party basis. Under this proposal, an amendment would only be valid if a majority in a majority (or a predetermined minimum number) of political parties represented in parliament supports the proposed reform. This approach shifts focus from previous proxies for broad political consensus – notably legislative supermajorities – to a more direct pursuit of broad consensus. By ensuring that amendments garner approval among a predetermined minimum number of parties beyond the ruling party or coalition, regardless of its dominance, the approach recognises political parties as the primary vehicles for representing the diversity of popular sentiments. This method ensures that no single party, regardless of its size at a particular moment, can unilaterally change the constitutional rules governing the democratic game. To borrow a football metaphor, the rules of the game should not be changed only by the game's giants. If they were, the rules would simply entrench the interests of those dominant teams. Similarly, in constitutional law, even a legitimate political giant should not have the exclusive authority to dictate the rules by which it and other political groups must abide.

It is important to recognise that the reliance on parties as a measure of the legitimacy of constitutional amendments is in itself a proxy for ensuring a broad consensus on the amendment among the people. However, by requiring a degree of support from individuals from different groups across the political spectrum, the proposed process allows us to conceive ‘the people’ in a pluralistic sense, reflecting the complexity of societal views more realistically. This proposal does not constrain majoritarianism; rather, it challenges us to rethink what constitutes a majority. Indeed, it defends majoritarianism, but one in which the majority for or against a constitutional amendment more accurately represents the popular will by encompassing voices from various contemporary political colours. While it is true that countries with proportional electoral systems tend to have more diverse parliaments, this diversity is no safeguard against inherent incumbency advantages and abuses. Therefore, specific rules requiring cross-party approval may be necessary to prevent any transient dominant majority from unilaterally altering the rules of the game. The fragmentation of parliament into multiple diverse groups also underscores the importance of finding consensus on fundamental constitutional provisions across these parties, which extends beyond supermajority requirements.

Cross-party approval requirements could be justified on both legitimacy and practical grounds. As amendments would require inter-party engagement and compromise, the process would be more likely to produce outcomes that approximate the popular sentiment, which is the principal determinant of democratic legitimacy. ⁸⁵ From a practical standpoint, the process offers several benefits. First, the process fosters deliberation and ensures that the passions, reasoning, ideologies, and interests of a diverse array of groups are understood and considered. In the African context, requiring inter-party approval for constitutional amendments could curb the almost reflexive resort to capricious and often self-serving constitutional amendments. Beyond preventing such amendments, the process would encourage a culture of regular engagement and compromise, thereby promoting understanding and mutual recognition of legitimacy between different groups. The experience of collaborative decision-making could engender a politics of trust and moderation, extending even into daily political discourse.

The proposal also offers a potentially more defensible and effective approach to protecting fundamental substantive principles than incorporating unamendable provisions or relying on judicial review of the substantive validity of constitutional amendments.⁸⁶ Unamendable provisions are often justified on hard-to-delineate normative or substantive grounds that may impose undue restrictions on contemporary polities and could be seen as limiting the people's right to self-governance.⁸⁷ By contrast, the elegant reliance on a process that fosters consensus among diverse political forces, representing a wide range of constituencies, interests, and ideas, may reduce the necessity or urgency of unamendable provisions.

Nevertheless, the cross-party approval requirement could be challenged on a number of grounds. First, it could make constitutional change more difficult, favouring the status quo and preserving the judgments of past generations at the expense of the contemporary generation's right to self-governance. This could result in a situation where the fear of illegitimate amendments sustains an illegitimate constitutional framework.⁸⁸ However, this criticism overlooks a crucial point: the proposed process would not arbitrarily obstruct change but rather ensure that change occurs only when there is genuine majoritarian will. In practice, as well, the African experience with abusive constitutional amendments provides a compelling rationale for establishing a mechanism that effectively slows down and, when necessary, prevents certain amendment efforts. Thus, the increased difficulty could be seen as an acceptable trade-off.

A related concern is that the proposed process might excessively empower smaller political groups. However, this concern must be contextualised within the African political landscape, where electoral processes often favour the incumbent, inflating the perceived level of support they command.⁸⁹ Moreover, to mitigate a potential ‘tyranny of the minority’, where numerous small parties could force constitutional change, any constitutional amendment must secure an absolute legislative majority, in addition to gaining support across parties.

Critics might also argue that relying on political parties to represent public views is elitist, especially at a time when representative democracy, political parties, and the broader ‘establishment’ are grappling with a crisis of legitimacy. This challenge is not unique to the proposed process, however, but also affects traditional amendment procedures. In direct comparison, the proposed cross-party consensus approach offers a more nuanced and representative measure of diverse popular preferences than supermajority or other traditional amendment rules. To address concerns about elitism and enhance legitimacy, the proposal could be complemented by direct public engagement mechanisms, including the requirement of popular referendums. This dual approach – combining cross-party political consensus with direct public input – could provide a more robust safeguard against self-serving or unpopular reforms, especially as the proposal has only covered a limited set of constitutional provisions. Referendum requirements, in particular, could empower opponents to constrain collusive efforts by political parties to advance narrow interests.⁹⁰ At the same time, referendums should not be seen as alternatives to the proposed amendment process, as they are vulnerable to manipulation and may not guarantee the genuine broad consensus necessary for legitimacy.⁹¹ Accordingly, referendums should only be deployed as a final step, following approval in cross-party forums.

The proposed amendment procedure is rare, but not without precedent. Under Thailand's 2017 Constitution, which was dominated by the military, constitutional amendments must be approved on final reading by an absolute majority in a joint sitting of the National Assembly, which consists of the House of Representatives (500 elected members) and the Senate (200 appointed members).⁹² In addition, the amendment must receive support from at least twenty percent of the members of opposition parties in the House of Representatives – defined as parties whose members do not hold the positions of Minister, President, or Vice-President of the House. Moreover, at least one-third of the total number of Senators must also approve the amendment. For certain types of amendments, the constitution further mandates approval through a referendum.

The Thai constitution in general, and its amendment provisions in particular, were crafted by the military regime that oversaw the constitution's drafting to prevent any single political group from easily pursuing constitutional amendments. This design was a response to the history of dominance by parties established by, or affiliated with, former Prime Minister Thaksin Shinawatra, which had won every election between 2001 and 2011.⁹³ Rather than imposing a high supermajority requirement, the constitution opted for a combination of a relatively low legislative majority and the necessity of support from opposition parties. This approach was seen as a more effective way of balancing the perceived need for reform with a broad political consensus on constitutional change.

The Thai amendment process can be seen as part of the military's strategy to consolidate its design choices, rooted in a deep distrust of electoral politics and political parties. This distrust is evident in the establishment of numerous powerful independent institutions, a non-partisan Senate, and mechanisms that guarantee the military's influence.⁹⁴ Combined with these powerful non-elected institutions, the amendment process is likely to make constitutional change particularly difficult. Nevertheless, despite their undemocratic origins, the amendment requirements are achievable and ensure that no single political player can unilaterally alter the constitutional framework. The consequence of a cross-party approval requirement is not so much to make change impossible, but to encourage a culture of compromise and engagement, which could ultimately benefit the democratic process and constitutionalism. Even if the process does make constitutional change more difficult, this difficulty may be justified given Thailand's history of political manipulation and dominance – a context that resonates with the experiences of many African countries.

Another country where constitutional amendments effectively require cross-party approval is Jamaica. Certain amendments must secure a two-thirds majority in both the House of Representatives and Senate.⁹⁵ As eight of the twenty-three Senators – about thirty-five percent – are appointed by the Governor-General on the advice of the Leader of the Opposition,⁹⁶ constitutional amendments to key provisions cannot be enacted without the support of at least one Senator nominated by the opposition leader, regardless of the ruling party's dominance in the House of Representatives. Although this sets a relatively low threshold for opposition concurrence, it still compels cross-party engagement and compromise. Similar provisions exist in other Caribbean constitutions, where Senators appointed on the advice of the opposition or other professional and business groups, acting together, can likewise necessitate that amendments receive support beyond the ruling parties.⁹⁷

In countries with politically salient linguistic, ethnic, and/or religious cleavages, constitutional amendments may also require cross-community approval, which in effect often coincides with cross-party approval. For example, Cyprus's now-defunct constitution required constitutional amendments to be approved by two-thirds of the elected representatives of the Turkish community, and by two-thirds of those of the Greek community.⁹⁸ In Africa, Burundi's 2018 Constitution guarantees the Tutsi minority forty percent of seats in the National Assembly.⁹⁹ Amendments require a four-fifth approval in the National Assembly and a two-thirds approval in the Senate,¹⁰⁰ effectively allowing the Tutsi minority to block amendments. While cross-community approval requirements can serve as good proxies for cross-party approval, this alignment is not always guaranteed. Dominant cross-ethnic parties could potentially push through amendments unilaterally. In Burundi, for instance, parties are prohibited from organising along ethnic lines, which means that each party must have membership from across communities. Accordingly, a dominant party with broad membership might already have sufficient membership to enact amendments. Additionally, Burundi's constitution allows the President to submit proposed constitutional amendments directly to a referendum¹⁰¹ without the need for legislative approval.¹⁰² This provision effectively undermines the consociational political arrangement and allows the President to bypass elite and political party consensus. Indeed, a similar provision in the 2005 constitution enabled the then President to propose the current 2018 constitution, which was ultimately approved by referendum.

The recognition of parties as centres of political power and agents of checks and balances extends beyond constitutional amendment procedures. Several constitutions require opposition involvement in key appointments, exemplifying this principle. The 1993 Constitution of Seychelles offers a notable example with its Constitutional Appointments Authority. The Authority is tasked with proposing candidates for key appointments and removals, including the Attorney General, members of the Electoral Commission, top judges, and the Ombudsperson.¹⁰³ The Authority is composed of five members: two appointed by the President, two by the Leader of Opposition, and a chairperson selected by these four members. To safeguard their independence, members serve guaranteed seven-year terms and can only be removed through strictly defined processes.

Similarly, Botswana's 1966 Constitution, as amended, mandates that five of the seven members of the Electoral Commission are selected by an All-Party Conference – a meeting of all registered parties convened by the Minister of Justice (the Judicial Service Commission selects the chairperson and one other member of the Commission).¹⁰⁴ In several other African countries, electoral commissions are similarly composed of representatives of the ruling and opposition parties, as well as neutrals. For instance, the Mozambican National Electoral Commission includes representatives of the ruling and opposition parties and civil society organisations, with the chairperson being a civil society representative.¹⁰⁵ The Electoral Commission of Côte d'Ivoire similarly includes representatives from the ruling and opposition parties.¹⁰⁶ Tunisia's 2014 Constitution goes a step further by giving the opposition the chairmanship of the parliamentary Finance Committee and the role of rapporteur for the External Relations Committee.¹⁰⁷ In particular, the opposition is granted the right to establish and lead an annual committee of inquiry. These examples illustrate a growing trend in African democracies to institutionalise opposition participation in critical governance structures and to promote sustainable systems of checks and balances.

Various Caribbean constitutions also grant prominent roles to opposition groups in key appointments.¹⁰⁸ In Jamaica, the Prime Minister must consult the Leader of the Opposition before appointing the Chief Justice, the President of the Court of Appeal, and three nominated members of the Judicial and Legal Service Commission, the Public Service Commission, and the Police Service Commission.¹⁰⁹ In Dominica, two of the five members of the Electoral Commission are appointed by the President, but only on the binding advice of the Leader of the Opposition.¹¹⁰ Similarly, in Antigua and Barbuda, the Leader of the Opposition nominates one of the four members of the Constituency Boundaries Commission and must be consulted by the Prime Minister in the nomination of the commission's chair.¹¹¹ Most strikingly, Caribbean constitutions typically grant the Leader of the Opposition the right to nominate a minority of the members of the Senate, ranging from ten percent of Senators in Barbados (Article 36(3)) to thirty-five percent of Senators in Jamaica (Article 35(3)). ¹¹²

While Thailand may appear to be an outlier in explicitly requiring cross-party approval for constitutional amendments, the examples discussed above demonstrate a growing recognition of political parties as crucial centres of checks and balances, and provide tangible support for proposals to reimagine constitutional amendment procedures. The approach proposed in this article draws on and extends these salient and practical approaches to safeguarding the integrity of democratic processes, to which the constitution is central.

Conclusion

Democratic backsliding – the slow but steady tampering of formal and conventional institutions to undermine democratic competition and the idea of limited government¹¹³ – has become one of the most pressing challenges of our time. Constitutional amendment procedures have become the tools of choice to convert transient electoral victories into long-term electoral advantages. Exacerbated by the global democratic recession, democratic progress in Africa – a continent starting from a low base – has largely stalled, with resurging coups d’état posing a significant threat.¹¹⁴ Yet, there are still reasons for optimism. Popular support for democracy remains consistently high across the continent;¹¹⁵ several African countries have demonstrated historical peaceful transitions of power to opposition parties in recent years, including Madagascar, Malawi, Zambia, the Democratic Republic of the Congo, Sierra Leone, and Liberia; and incumbents in Niger, Nigeria, Mauritania, Botswana, and Burundi have respected term limits.

At the same time, Africa remains a continent with some of the longest-serving presidents, who continue to resort to constitutional and legal amendments to consolidate their power, undermine the opposition, and, more broadly, erode the democratic dispensation. In settings where dominant political groups prevail, existing constitutional amendment requirements have proven too easy to meet. In such practically non-competitive regimes, the traditional safeguards of legislative supermajority and referendums have largely failed to prevent regressive, self-serving amendments.

With these problems in mind, this article proposes a novel approach to amendment procedures, aimed at curbing abusive constitutionalism and strengthening democratic institutions. The core argument is that requiring cross-party approval for constitutional amendments can serve as a more effective safeguard against the erosion of democratic ideas and institutions in contexts of dominant parties, and strengthens emerging practices of deliberation and compromise in places with increasingly competitive electoral outcomes. While this approach admittedly makes constitutional change more difficult, thereby favouring a status quo that may be neither workable nor even neutral (thus compromising ideals of self-governance), it would apply only to a limited set of the most fundamental rules of constitutional democracy, and the envisaged difficulty may be precisely what is needed where change has been too easy. Given the experience with defective democratic regimes in many African contexts, where government-initiated reforms often do not equate to genuine self-governance, the proposal presents a reasonable trade-off.

The article does not claim that the proposal is a panacea, bulletproof, or an impenetrable defence against authoritarian leaders. However, its implementation would at least have forced leaders to either make concessions to the opposition, go to the lengths of forming sham opposition parties, or bypass the constitutional framework altogether. Any of these actions would strip them of the moral high ground of claiming constitutional legality. Traditional constitutional amendment processes, even when stringent, can still lend an air of legitimacy to autocratic changes. Constitutions alone cannot halt the tide of dictatorship. But neither should they legitimise it.

The proposed amendment process is also normatively defensible. By ensuring support across major political leanings, it more genuinely reflects the broad popular consensus that must underlie the basic rules of a democratic dispensation. The proposal does not seek to preclude the tyranny of the majority, only the tyranny of a single party or a faction. Nor does it oppose majoritarianism, thus avoiding potential claims of undermining popular self-governance. Instead, it encourages us to reconsider what constitutes a majority in order to accurately reflect broad consensus.

The article lacks specificity as to the precise operationalisation of the proposal. The procedure could vary from requiring a supermajority approval by a supermajority of political groups represented in parliament to a majority approval in most, or a defined minimum number, of parties, while ensuring that the total supportive votes in the legislature constitute an absolute majority. The potential combinations of majority thresholds within each party and the number of required parties are virtually limitless, constrained only by human imagination. In any case, the rules would have to be tailored to the specific context in which they are applied.

Finally, this proposal must be presented with a caveat. The surest way to protect and sustain a democratic dispensation is to create the conditions for a genuinely competitive democracy and a space for free expression of the people, civil society, and the courts. While the focus of this article has been on political parties, the ultimate survival of democratic constitutionalism and a political system grounded in freedom hinges on the vanguard of these groups. Therefore, while the proposal can help check attempts to undercut abusive constitutional changes, it is not a substitute for reforms aimed at limiting the overbearing presidency, innovative mechanisms to address winner-takes-all politics and the incumbency advantage, strengthening the capacity, resources, and policy orientation of political parties, and deepening popular understanding of, and support for, constitutional democracy.