A crisis of design and judicial practice? Curbing state disengagement from the African Court on Human and Peoples’ Rights

In the space of four years, between 2016 and 2020, four of the ten states that had recognised the jurisdiction of the African Court on Human and Peoples’ Rights to receive cases directly from individuals and NGOs withdrew their declarations made under article 34(6) of the Court Protocol. While this form of contestation is not unprecedented in the history of states’ behaviour towards international courts, this article argues that the disengagement from the African Court’s jurisdiction involves peculiarities that specifically relate to the Court’s system design and its practice. The main contention in the article is that the declarationbased adherence to the African Court’s jurisdiction is in crisis due to a costbenefit imbalance. The article argues that although all four withdrawals resulted from decisions of the Court on important and contentious domestic socio-political issues, systemic features such as the lack of appeal, an overly restrictive review mechanism and the weak functioning


Summary:
In the space of four years, between 2016 and 2020, four of the ten states that had recognised the jurisdiction of the African Court on Human and Peoples' Rights to receive cases directly from individuals and NGOs withdrew their declarations made under article 34(6) of the Court Protocol. While this form of contestation is not unprecedented in the history of states' behaviour towards international courts, this article argues that the disengagement from the African Court's jurisdiction involves peculiarities that specifically relate to the Court's system design and its practice. The main contention in the article is that the declarationbased adherence to the African Court's jurisdiction is in crisis due to a costbenefit imbalance. The article argues that although all four withdrawals resulted from decisions of the Court on important and contentious domestic socio-political issues, systemic features such as the lack of appeal, an overly restrictive review mechanism and the weak functioning * ENA, Licence (Benin) LLM LLD (Pretoria); sathorace@yahoo.fr. The opinions expressed in this article are the author's own and do not purport to reflect the opinions or views of the African Court on Human and Peoples' Rights. I am grateful to the Journal for editorial guidance, and to Ms Lee P Angel for her support in preparing this article.

Introduction
The African Court on Human and Peoples' Rights (African Court), amid euphoria and expectation, was established as a fully-fledged judicial organ. 1 Yet, the Court arguably was born with congenital defects. In designing the Court, African Union (AU) member states put in place a two-tiered system of state adherence. Ratification of the Protocol to the African Charter on Human and Peoples' Rights on the Establishment of an African Court on Human and Peoples' Rights (African Court Protocol) presents the first tier of acceptance, allowing for indirect access to the Court via the African Commission on Human and Peoples' Rights (African Commission). As at July 2020, 30 of the 55 member states of the AU had ratified the Protocol. Depositing a declaration under article 34(6) of the Protocol, thereby accepting direct access to the Court of individuals and non-governmental organisations (NGOs) enjoying observer status with the Commission, provides for the second tier of acceptance. 2 Only ten states, with Burkina Faso being the first in 1998 and The Gambia the last to do so in 2020, had ever adhered to both tiers. 3 As the Court's operations intensified, the second tier has come to threaten the very operational viability of the institution. In the space of four years, between 2016 and 2020, four of the ten states that had previously filed it, withdrew their article 34(6) declaration: Rwanda in 2016, Tanzania in 2019, and Benin and Côte d'Ivoire in 2020. Between them, the four withdrawing states have approximately 85 per cent of the litigation before the Court. The remaining 15 per cent of the Court's cases were brought against states that are either recently restored democracies, such as The Gambia and Tunisia, or post-transitional democracies grappling with terrorism or other forms of instability, such as Burkina Faso and Mali. Notably, none of the AU's 'big five' -Algeria, Egypt, Morocco, Nigeria and South Africa -has filed the declaration, 4 nor has Ethiopia, which hosts the AU headquarters. It is beyond dispute that this emergent trend of withdrawal jeopardises the Court's mandate to protect human rights on the African continent. 5 State contestation is not unique to the African Court. Rather, it reflects a global trend, 6  In examining the problem, labelling states as rogue and antirule of law may be attractive, 10 but tells only part of the story. It is important, instead, to appreciate why states join an international regime such as the African human rights system. Theories explaining state adherence to international law and mechanisms include coercion, persuasion and acculturation. 11 With a specific emphasis on the benefits of adhering to international adjudication mechanisms, commentators have argued that states join despite their distrust in the systems because they balance the costs and benefits to them, and set up shielding mechanisms such as the requirement to make specific declarations recognising a body's jurisdiction or accepting individual access. 12 Accordingly, as to what motivates states to confront or disengage from international courts, empirical work has identified the cost of membership, the influence on the domestic political system, and the domestic impact of judgments. 13 Ultimately, when the cost of adhering is significantly higher than the benefits, states may activate the shield by withdrawing. 14 The current status of withdrawals from the jurisdiction of the African Court attests to the fact that states' trust in the Court is on the decline. 15 In the wake of this trend, the all-important question is how the challenge to the Court's legitimacy should be addressed. In this article I advance the argument that the growing discontent among states must be addressed and redressed in a systematic manner with a view to curbing withdrawals and rebuilding state adherence. 10  I begin the discussion by exploring the African Court's current jurisdictional crisis, focusing on the factors that precipitated the withdrawals. The discussion proceeds to demonstrate why and how the design of the system, the practice of the Court and the weak functioning of shielding mechanisms, including the African Commission, ignited a fast-tracked process of diminishing state cooperation. Finally, I expand on some of the solutions that in my view are most likely to help reinvigorate the Court's legitimacy and restore states' confidence in the system.

A jurisdictional and existential crisis
I argue that the Court faces a crisis that is jurisdictional and existential in nature, as the current state of affairs has a critical impact not only on the scope of intervention of the Court but also on its authority and legitimacy. The crisis is one of both substance and scope. Below, I analyse the language used as part of states' contestation, and consider the Court's response, while assessing the validity of the arguments arising from these statements.
The African Court faces a crisis in substance. The extent of withdrawal is a crisis as withdrawals are among the worst forms of contestation in the history of state-led challenges to international courts. 16 Most severe of all would be for states to denounce the African Court Protocol, or not to renew the terms of judges, as has occurred in respect of the SADC Tribunal. 17 Although it is not a statutory requirement in the Court Protocol that states should reveal the reasons for either making or withdrawing their article 34(6) declarations, all four states that withdrew have given some indication of their motivations. The discussion therefore relies on the language of contestation as a relevant prism through which to fully appreciate the nature of state discontent. In the analysis it is highlighted that all four withdrawals had remote or proxy grounds in the Court's rulings pertaining to socio-political issues that are of critical importance in the national sphere. 16

Rwanda's abrupt withdrawal over a perception of Court partiality and matters related to genocide
Rwanda filed its declaration on 22 January 2013. When withdrawing it on 24 February 2016, Rwanda stated that it was doing so over 'a fugitive from -the Tutsi genocide -justice [who] has, pursuant to the above-mentioned Declaration, secured the right to be heard by the Honourable Court, ultimately gaining a platform for reinvention '. 18 Subsequently, on 30 January 2017 the government of Rwanda informed the registry that it would no longer participate in proceedings before the Court on the grounds that the process with regard to cases involving Rwanda was not independent; that its outcome was pre-determined; and because the Court refused to hold a hearing in the matter of Victoire Ingabire Umuhoza v Rwanda 19 to rule on the issue whether it was proper for genocide accused or convicts to be granted access. 20 The state also objected to the fact that the Court, without informing the parties, issued a substantive corrigendum signed only by the Court's President and not the full bench, as was the case with the initial decision. 21 The Court formally responded to Rwanda's contestations through two channels. Regarding the withdrawal, the Court gave a judicial response through its ruling on jurisdiction of 3 June 2016 in the Umuhoza case. 22 There, it ruled that the withdrawal was valid, would take effect after 12 months of the deposit of the withdrawal with the AU, and did not affect pending cases. On the administrative side, the registry of the Court informed the state that its cooperation with nongovernmental institutions should not be read as an encroachment on the independence of the Court, and that Rwanda was free to raise the issue in the course of the proceedings. With respect to nonparticipation, the registry also communicated to the state that the service of pleadings on both parties would proceed in adherence with the relevant rules. 23 I argue that the language used in Rwanda's communication is one of crisis, as it reveals an attack on the independence and legitimacy of the African Court. Whether the reasons advanced are legally tenable is another issue altogether. Allowing alleged genocide fugitives to file cases before the Court cannot in itself warrant a conclusion on the impartiality of the Court. A judicial institution operates based on statutes and neither the Court Protocol nor its Rules provide for the status of genocide convict as one of the criteria governing jurisdiction or admissibility. 24 The Court's response to the state's decision to cease participation is also legally sound given that service, the exchange of pleadings and default judgment were in accordance with statutory prescripts. The true reason for the withdrawal may well be, as some commentators suggest, that Rwexit -Rwanda's withdrawal of its declaration -sought to prevent the Court from revealing the state of socio-political governance in Rwanda. 25 Nevertheless, perhaps the Court should have granted the hearing on the question of whether suspects or convicts of crimes as grave as genocide should be allowed to submit a case to the Court. By doing so it would have used the opportunity to make a ruling on an interesting and important question at the intersection between international criminal and human rights law.

Tanzania's exit over litigation fatigue coated in justification based on reservation
The second state to withdraw its declaration was Tanzania The issue of the reservations and their validity formally arose for the first time when Tanzania adduced them as a ground for withdrawal. While international law allows states to make reservations to treaties, 28 or recognition of jurisdiction, 29 it is generally accepted that these limitations should not hamper the discretionary exercise by the tribunal of its jurisdiction. 30 As such, reservations should be clear, and not 'fake' or 'illicit'. 31 'Fake' reservations are those that are superfluous because they provide for an exception that is inherent in the applicable law. 32 Against this understanding, Tanzania's reservation relating to the exhaustion of local remedies would not stand the test of validity as the reserved rule is already enshrined in the applicable law before the Court as a condition of admissibility under article 56(5) of the Charter, and Rule 40(5) of Court Rules. 33 'Illicit' reservations are those that mainly touch on material jurisdiction. 34 Reservations on subjective national jurisdiction, for instance, fall under this category when they empty the declaration of its object and purpose. 35 Tanzania's second reservation that direct access for individuals and NGOs 'should only be granted … in adherence to the Constitution' exemplifies an 'illicit' reservation. Such reservation is not in conformity with articles 5(3) and 34(6) of the African Court Protocol, which provide for the filing of the declaration as the sole condition of access. The reservation is 'illicit' in the sense that it annihilates the very purpose of the declaration, which is to allow direct individual access to the Court, including challenging the conformity of the Constitution with international law ratified by the concerned 2020 Tanzania's withdrawal came as the conclusion to an incremental contestation process, was not abrupt and could, therefore, have been foreseen and prevented. Some of the most systematic and consistent acts of contestation include objections to the African Court exercising both first instance and appellate jurisdiction, 38 and overstepping the authority of apex municipal courts on issues such as nationality 39 and the death penalty in respect of which the state consistently affirmed that sentences were valid under international law. 40 Prior to withdrawing its declaration Tanzania also took steps related to cooperation with the Court, including by seeking an extension of time in individual cases and initiating or participating in meetings between the registry, the Attorney-General and the Solicitor-General to discuss more effective judicial cooperation. 41 There is no evidence that the proposals towards an improved judicial 36  It could be anticipated from the above engagements that Tanzania was not prepared to carry a technically-challenging and increasing international litigation load. The state's long-standing and consistent challenge to meet deadlines in dozens of competing proceedings only resulted in an ultimate attempt to face the burden by resorting to a wholesale extension of time. 42 Seemingly, while the Court acceded largely to grant the requested extensions within the feasible time frames, Tanzania had reached litigation fatigue.
It is worth recalling that, at the time of withdrawal, Tanzania had been the defendant in 138 out of a total of 255 cases received by the Court. 43 In addition to dealing with such litigation load, Tanzania has to carry the burden of implementation of over 60 administrative, legislative, judicial and pecuniary orders resulting from the Court's judgments. Most notably, when since 2016 reparations proceedings began to mature, the total damages in all resolved cases reached TSh 154 110 000 (approximately US$ 106 618). Considered in isolation, this figure may be relatively affordable to a state such as Tanzania. However, the perception might be different when the figure is examined in the context of the legitimacy debate. It is worth noting that Tanzania had to honour these orders, concurrently and cumulatively. At the time of withdrawal four of the non-pecuniary orders had been implemented, while the cumulated time allocated for implementation was three years for the pecuniary orders and four years for other types of orders African Court Registry, List of cases received as at 13 July 2020. As at 13 July 2020 the state had been dealing with 143 of the 285 applications received since inception. At the same date it was still dealing with 108 pending cases, which is more than half of the total cases pending before the Court against all ten states that filed the declaration. 44 African Court Activity Report 2019, https://en.african-court.org/images/Activity %20Reports/EN%20-%20EX%20CL%201204%20AFCHPR%20ACTIVITY%20 REPORT%20JANUARY%20-%20DECEMBER%202019.pdf (accessed 19 July 2020).
All judgments of the Court being final, it seems that in the circumstances no option other than withdrawal would have been the reasonable course of action for a state concerned. This account does not deny that withdrawing indisputably represents a setback for justice and rights protection guaranteed in the African Charter.
Although neither the notice of withdrawal nor an official statement mentioned it, there is reason to believe that the Court's ruling on sensitive issues of socio-political relevance in Tanzania may have contributed significantly to the withdrawal. Following its first ever judgment on the merits delivered in the Mtikila case, ordering Tanzania to amend its Constitution and allow independent candidatures, the African Court made several rulings that touch on the operation of the judiciary in Tanzania mainly with respect to fair trial rights. 45 Furthermore, in the matter of Robert John Penessis v Tanzania the Court ruled that Tanzania had violated the applicant's right to nationality in proceedings where all evidentiary options arguably had not been investigated, as discussed more fully later in the article. 46 In the case of Ally Rajabu & Others v Tanzania the Court also ruled that the provisions of the Criminal Code prescribing the mandatory death sentence in cases of murder violated the right to due process and life, and ordered that the Code be amended to remove the sentence. 47 With respect to the core operation of the domestic justice system, the Court has consistently held that the review of judgments and constitutional petitions for breach of fundamental rights are extraordinary remedies that an applicant is not compelled to exhaust under article 56(5) of the African Charter. 48 Besides, some of its leading rulings such as in the Thomas, 49 Abubakari 50 and Onyango 51 cases may have been perceived as delegitimising domestic criminal policy, in that they indirectly reversed the rulings of the highest 45 See A Possi '"It is better that ten guilty persons escape than that one innocent suffer": The African Court on Human and Peoples' Rights and fair trial rights in Tanzania

Benin's withdrawal for economic interests and regime stability
Benin filed its declaration on 8 February 2016. The African Court received the first application involving this state on 27 February 2017, which was also the only case for the year 2017. As at 13 July 2020 the Court had received a total of 28 cases involving Benin, more than half filed by or about political opponents including matters related to the May 2020 local elections and the upcoming March 2021 presidential election. 53 In the notice of withdrawal dated 24 March 2020 the Benin government stated that it withdrew the declaration because the Court implemented it in a manner that was 'perceived as a licence to interfere with matters that escape its competence causing serious disturbance to the municipal legal order and legal uncertainty that is fully detrimental to the necessary economic attractiveness of State Parties'. 54 As '[o]ne of the regrettable interferences and annoying disturbances', specific mention is made of an order of the Court suspending the enforcement of a domestic court judgment for seizure of property to honour a bank loan in a commercial deal between private persons.
Subsequent to the filing of the notice of withdrawal, the Minister of Justice further explained Benin's decision to withdraw as the consequence of 'several inconsistencies in decisions rendered by the Court in the recent years that have led Rwanda and Tanzania to withdraw their declarations'. 55 He justified Benin's withdrawal by 52  the fact that it was 'impossible to sanction -these inconsistencieswhich the Court itself has not given the impression that it is keen on addressing'. 56 The Minister concluded the statement by reiterating that Benin remained a party to the Court Protocol and announcing that 'the President has decided to present to his peers a reform of the judicial institutions aiming at accelerating the operation of an African Court of Justice and Human Rights as contemplated during the 3rd ordinary session of the Assembly … in July 2004'. 57 The order referred to in the notice was that for provisional measures issued in the matter of Ghaby Kodeih v Benin. 58 In this ruling the Court directed Benin to suspend the transfer of the property deed to the creditor of the domestic court judgment in the Kodeih matter, as well as any measure of dispossession of the applicant. 59 It is of critical importance to first deal with the suggestion in the notice that the Court overstepped its powers by asserting jurisdiction in a matter that the state averred was commercial in nature, and thus beyond the Court's jurisdictional purview as a human rights court. In my opinion, disputes that were originally commercial, criminal or labour law-related, or even inherently administrative or constitutional, may end up being catalogued as human rights matters as long as they involve violations of human rights contained in a relevant instrument. 60 As examined by domestic courts, the Kodeih matter is related to commercial and business law, and arose from a dispute between private entities. However, in the case before the African Court the applicant alleged the violation of his right to appeal on the ground that the domestic lower court issued a judgment that was rendered as final and proceeded to authorise the enforcement thereof. The applicant averred that doing so also violated his right to property, as execution of the judgment would involve transferring the deed of the property to the creditor. That such allegations fall within the jurisdiction of the African Court is beyond dispute. The issue may be elsewhere, namely, that the Court would not have exercised actualite/635/retrait-benin-cadhp---declaration-ministre-justice-legislation/ (accessed 12 May 2020). 56 As above. jurisdiction while the matter was obviously still pending final adjudication in the domestic system and, therefore, it inevitably would have been found inadmissible for not exhausting domestic remedies.
The fact that the notice of withdrawal mentions the Kodeih matter only as 'one of the instances of interference' suggests that Benin had further reasons for discontent and disengagement. As a matter of fact, between November 2018 and April 2020 the Court issued no less than eight 'critical' decisions against Benin, most of them involving political opposition figures.
On 5 and 7 December 2018 the Court issued twin orders in the matter of Sébastien Germain Ajavon (CRIET) v Benin, the first reopening pleadings to receive further submissions by the applicant, 61 and the second directing Benin to suspend the execution of a 20-year prison sentence for drug trafficking. 62 In its report to the Court on implementation of the second ruling the state argued the material impossibility to enforce the order on the ground that the African Court had issued it ultra vires, and that the decision was in breach of its sovereignty. 63 The argument made earlier in the Kodeih matter applies here: The Court's ruling cannot be challenged on jurisdiction as it acted within the Court Protocol and its Rules of Procedure. Be that as it may, Benin's report to the Court already indicated a posture of defiance and hence foretold a looming crisis. The following decisions appeared to have turned the looming crisis into direct confrontation and, finally, into divorce.
On 28 March 2019 the Court delivered its judgment on the merits in the Ajavon CRIET case, where it found several violations, mainly related to the right to appeal, and ordered Benin to take measures to annul the 20-year sentence and amend its laws to allow appeal of the rulings of the Court for Economic Crimes and Terrorism (CRIET). On 5 February 2020 Parliament adopted an executive amendment Bill establishing an Appeals Chamber within the CRIET 64 in implementation of the judgment. 65  In its judgment on reparations delivered on 28 November 2019 the African Court ordered various pecuniary reparations reaching an award totalling an unprecedented US$ 66 million to the applicant for both material and moral damages. 67 While Benin did not express any official contestation to either the merits or reparation judgments, its submissions in both proceedings were that the Court lacked jurisdiction and that the matter was inadmissible. 68 With respect to admissibility, it appears that the applicant made no attempt to approach the Constitutional Court of Benin with a plea for the violation of his rights in the proceedings before the CRIET. 69 Less than three months after issuing the order for payment of US$ 66 million in the Ajavon CRIET reparation judgment, the African Court on 28 February 2020 issued the order in the Kodeih case, as discussed earlier. Benin filed the notice of withdrawal less than a month later, on 24 March 2020. The last decision in the row is the order issued by the Court on 18 April 2020 in the Ajavon -Local Elections matter. 70 The applicant's case is that a set of new laws enacted in preparation of the 17 May 2020 local elections in Benin violated his right to political participation and generally constituted a setback to democracy and a breach of the Beninese peoples' right to elect their representatives. The Court ruled -exactly as the applicant had requested -that the elections should be suspended until it disposes of the merits of the case.
While this decision was issued subsequent to the article 34(6) withdrawal, there is evidence that the ruling exacerbated the state's contestation that the Court was overstepping its powers. As a matter of fact, the Minister of Communication stated at a press conference on 23 April 2020 that 'safeguarding the rights of a Benin national cannot override the normal functioning of our institutions. We will therefore be going to the poll comes 17 May.' 71 According to him, 66 See TC Nagnonhou 'Chambre des appels de la Criet: La Cour suprême installe le président' La Nation, https://lanationbenin.info/chambre-des-appels-de-la-crietla-cour-supreme-installe-le-president/ (accessed 11 May 2020 In both the substantive perspective and that of good administration of justice, the order on the election poses serious problems to which I return later when discussing the systemic approach to understanding the jurisdictional crisis facing the African Court. Suffices at this point to observe that the Kodeih order is only one of the many reasons that are invoked to justify withdrawal in a more general bid to sustain regime stability. Benin's withdrawal can reasonably be seen as a preventive shield against a too intrusive Court that may as well, for instance, go as far as threatening the incumbent President's bid for a second term by reversing the outcome of the March 2021 presidential election should the case arise. As evidence to this hypothesis, there is a perceptible trend, beyond the scope of the African Court, to evade human rights accountability through rule of law autarchy or isolationism. Arguably, as the most revealing illustration, the Constitutional Court of Benin ruled in a decision dated 30 April 2020 that the enforcement of the ECOWAS Court Supplementary Protocol, which had entered into force upon signature under a provisional treaty clause and rulings made thereunder, was void for lack of ratification endorsement by the National Assembly. 73 Technically viewed, the ruling amounts to a withdrawal from the jurisdiction of the ECOWAS Court, for which the applicable statutes do not provide.

Côte d'Ivoire's unexpected but foreseeable exit based on pure politics
Of the four withdrawing states, Côte d'Ivoire had the least turbulent relationship with the African Court prior to withdrawing its declaration, which it filed on 23 July 2013.  80 The applicants alleged that their detention violated their rights to a fair trial, and that the warrant was aimed at preventing Soro from standing as a candidate in the October 2020 presidential election. 81

Understanding the crisis with reference to design and practice
In this part I argue that the crisis facing the African Court's justice system, on the one hand, is due to an imbalanced, inadequate and incomplete structural design and, on the other, related to the role of the Court and other stakeholders in managing the system.

The declaration as shield
In my view, article 34(6) is a pertinent illustration of the weakness of the African Court justice system. The drafting history of the Court Protocol reveals that article 34(6), which conditions direct individual access on the filing of a declaration, did not form part of the text until the last stage of the process. 82 Several unsuccessful attempts made by the Court itself to have the declaration repealed are reflective of its awareness that the declaration represents a hurdle to its effective operation. 83 Attempts to have the declaration removed also came from outside the Court, as exemplified by a lone practitioner's submission of the case Femi Falana v African Union, which the Court dismissed on the basis that it lacked jurisdiction over the AU. 84 The Court might have been offered another opportunity to rule on the issue, albeit indirectly, in the recently-filed case of Glory Cyriaque Hossou and Landry Angelo Adelakoun v Benin, in which the applicants contend that Benin has violated the right of access to justice in the African Charter for withdrawing its declaration, and seek an order suspending the withdrawal. 85 The rationale behind these attempts to remove the declaration is that it renders the Court's remedial competence illusory. 86 However, even if article 34(6) detracts from the full potential of the Court, a plea for repealing it is overly optimistic and actually not advisable in the current context. I expound on the issue later when discussing how to enhance state adherence and oversight.

Lack of appeal or meaningful review
Pursuant to article 28(2) of the African Court Protocol, judgments of the Court are final. Therefore, there is no room for appeal whether for material or substantive errors -even where the Court would manifestly be misguided. Given the increasing trend of the Court to adjudicate on critical issues such as electoral disputes and awarding substantial damages, the right of states to appeal may make a significant difference to the likelihood of further article 34(6) withdrawals. 87 Without the possibility of recourse to a judicial appeal, states may well find disengagement the most appropriate means of protecting their sovereignty against an unfavourable ruling.
The procedure for review of decisions delivered by the African Court should also be explored as a factor contributing to the current withdrawal crisis. Article 28(3) of the Court Protocol appears to have tied the hands of the Court as reflected in Rule 67(1) which allows review only 'in the event of discovery of evidence, which was not within the knowledge of the party at the time the judgment was delivered'. However, the Court in designing its 2010 Rules arguably has underused the leeway of the Protocol that review shall be governed by 'conditions to be set out in the Rules'. The Court did not make use of the possibility to extend review cases from the traditional discovery of new evidence to a much more contextualised review for errors on material crafting, facts, law and legal interpretation. It instead took a literal and, therefore, very restrictive approach to the exercise of its discretion. 88 The standards of review remain unchanged under the new (2020) Rules, which entered into force on 26 June 2020.

Ineffective institutional shields
In weak or nascent international rule of law environments such as that of the African Court, a two-tiered structure -for example, with a quasi-judicial and judicial tier -may be needed to preserve the judicial tier from attacks before it has reached a certain level of maturity. Given its dialogical relationship with states, the African Commission should have played the role of filtering cases and shielding the Court from applications destined to cause major contestation.
Unfortunately, it appears that challenges relating to leadership and institutional preservation did not allow an effective adjudicatory complementarity as would have benefited the African human rights system. 89 For instance, the Commission declined to examine the four cases that it received from the Court, apparently on the ground that these cases were transferred by judicial rulings rather than through administrative channels. In ten years of complementarity 90 the African Commission also submitted only three cases to the Court. 91 As another consequence, the system did not take advantage of the fact that states are very likely to view institutional litigation by the Commission as less personalised, and thus more neutral, than proceedings generated by national political stakeholders or even civil society organisations. 88  The African Commission might also have been reluctant to file applications before the African Court for fear that the latter would examine the cases de novo, thus reopening issues already settled by the Commission. 92 This fear arguably led the Commission never to invoke or apply Rule 118(1) of its 2010 Rules that governed the submission of decided cases to the Court. 93 Similarly, the two institutions did not show solidarity in complementarity when the Commission faced threats from the policy organs of the AU, or when the AU policy organs demanded that the Court retract the names of non-compliant states from its Activity Report. 94 The lack of adequate peer engagement among AU member states within the policy organs may also have sent the message that contestation would occasion little or no cost. For instance, the AU Executive Council did not face significant opposition from any AU member state when it barred the Court from mentioning noncompliant states by name in its Activity Report, despite the clear mandate to do so in article 31 of the Court Protocol. The Council also allowed a practice of challenging non-compliance reports by states that did not file the declaration and thus were not involved in any of the cases under consideration. 95 This situation is compounded by the minimalist monitoring mechanism established under article 29 of the African Court Protocol, whereby a mere call to comply stands as sanction for non-compliance. 96

Shortfalls in the African Court's practice
The deleterious effects of poor design may be ameliorated by a wellcoordinated and purposive practice. As I argue below, a review of the African Court's practice reveals that regrettably it has not consistently sought to address or succeeded in addressing the design weaknesses of the system. 92

Problematic timing of adjudication
The limited acceptance by states of optional direct individual access placed the burden of litigating cases on a small number of states. Faced with a significant number of cases in respect of a single state, the Court may well from time to time have overlooked whether a specific decision was the most appropriate at the relevant time and within the particular institutional context.
Because the four withdrawing states have been litigating approximately 85 per cent of the total cases filed before the Court up to date, the Court from time to time had to issue several decisions against the same state within a short period, without necessarily giving proper thought to whether it was doing so at the most suitable time and in the best sequence. For instance, in the same year (2019), the Court delivered against Tanzania, a state already heavily burdened by remedial orders, a judgment in the Rajabu case outlawing the mandatory death penalty in cases of murder and a ruling effectively recognising the Tanzanian nationality of the applicant in the Penessis case. As stated earlier, the death penalty and nationality issues have deep social and political resonance in the country. In respect of Benin, the question may similarly be asked whether the Kodeih and Ajavon Local Elections orders should have been delivered at the particular time, and whether the two decisions should have been handed down at a two-month interval.
The issue of when the African Court delivers its rulings could also be relevant to whether seeking justice in the Court worsens the situation of an applicant. For instance, the decision to grant or deliver the Soro order apparently had the consequence of bringing forward former Speaker Soro's trial. An international arrest warrant by then had been pending for five months, with no trial date set. Almost immediately after the order had been delivered on 22 April 2020, a one-day trial was announced and held on 28 April 2020. Convicted of embezzlement and money laundering, the applicant was sentenced to 20 years' imprisonment, was ordered to pay a fine of €7 million, and was barred from civic duties for five years. Similarly, the Ajavon Local Elections order might have stood a better chance of being implemented had it been delivered several months prior to the elections, and had it covered only the situation of the applicant as was done in the Houngue case, discussed later. 97 97 Houngue Eric Noudehouenou v Benin, AfCHPR (Provisional Measures, 6 May 2020).
The momentum of adjudication features critically in the Ajavon Local Elections case, referred to earlier. As a general rule, election adjudication inherently is of domestic and, more specifically, constitutional jurisdiction. According to practice, supreme or constitutional courts are time bound by electoral law, including the national constitution. It therefore is of critical importance for an international court vested with a related mandate to be conscious of these standards. In this case the Court received the application on 29 November 2019, and on 9 January 2020 received the specific request for provisional measures in respect of the elections to be held on 17 May 2020. 98 The Court issued the order suspending the elections on 17 April 2020, 30 days before the scheduled polling date.
In this instance, also, one cannot be blind to the contradictions in the Court's procedural practices. In its established practice the Court, in the interests of justice, had shortened time for the submission of pleadings or the implementation of orders. 99 Besides, it had precedent-based power to issue orders suo motu where the circumstances so require, as it did in the African Commission (Libya Arab Spring) case. 100 Yet, in the Ajavon Local Elections case the Court allowed exchanges of pleadings and processed an evidently urgent electoral matter for more than three months, only to suspend the elections 30 days before the scheduled election date. A similar trend is observed in the cases of Jebra Kambole v Tanzania, involving the constitutional ouster of the result of the presidential election from the Court's jurisdiction, 101 and Suy Bi Gohore Emile and Others v Côte d'Ivoire challenging the composition of the Ivorian electoral commission. 102 In these two cases the Court delivered judgment on 15 July 2020, with forthcoming elections scheduled for October 2020 in both countries. 103 The Court's ability to be deliberate about the timing of its judgments is impaired by the non-permanent nature of the Court. With the exception of the President, all judges serve part-time. The Court's part-time nature allows for only four annual sessions of four weeks each. Given this factor and going by practice, the likelihood is slim of the Court ruling within a year on the merits of any case submitted to it. Suspending elections in April until the merits are decided, much later, as was done in the Ajavon Local Elections case, therefore amounts to postponing the election sine die with severe consequences for the country and government, including socioeconomic and political upheaval. As discussed later in the article, the Court had fairer and more contextualised alternative options.
The Court's Order in the Ajavon Local Elections case also constituted an ultra petita remedy, which lacked purpose, context and fairness. The Court ordered Benin to provisionally suspend the election in order to uphold the rights of a single applicant without balancing all other interests involved. While the applicant made a request for suspension of the poll based on alleged violations of his own right and that of the citizens of Benin, he adduced no evidence to speak on behalf of the entire people of the country in elections involving over five million voters. As he is not an institution vested with civil society mission, the applicant also lacked standing to bring a public interest case. In this context, the balancing of rights arose prominently in the sense that the Court should have assessed the interests of the applicant and possibly those of his political party 104 against the interests of the rest of the country. 105 As the main normative reference of the

Inconsistent assessment of evidence
The African Court's practice with respect to assessing evidence may also have contributed to loss of confidence by the states that withdrew and other states generally. As discussed earlier, the Court did not properly consider submissions of the respondent regarding the effectiveness of the Constitutional Court as a remedy to exhaust in the Ajavon CRIET matter involving Benin. I consider two more cases. In the case of Nguza Viking & Another v Tanzania the Court held that a determination on evidence for identification falls within the exclusive preserve of national courts to which the African Court defers, unless admitting that the evidence would lead to a miscarriage of justice. 109 The Court thus dismissed the second applicant's alibi on the ground that domestic courts had rejected this evidence. 110 Yet, it found a breach of right due to the domestic court's refusal to allow the impotence test requested by the applicant in a rape case. 111 The Court further affirmed the identification of the applicants, which domestic courts conducted during the trial by asking the applicants to shift seats whereas the victims had seen them twice prior to the identification during trial. 112 The assessment of evidence also arises in the Penessis case where the Court factually established the applicant's citizenship based almost solely on a copy of his birth certificate. 113 The respondent adduced extensive evidence, including copies of two passports (British and South African) bearing different names which the applicant used to seek entry into Tanzania. 114 The applicant never justified the difference in the names on the two passports, which featured in the decisions of domestic courts that he submitted to the African Court in arguing the admissibility of his application. 115 The Court's approach in these instances can only leave the impression that it does not give due consideration to submissions and evidence of respondent states. Decisions of the Court therefore may cause a sentiment of bias or unfairness to which states believe withdrawal is the effective response for lack of alternative means of contestation.

Inconsistent and incomplete judicial restraint in respect of admissibility
The admissibility-related law-making standards of the African Court are also questionable in relation to how it observes judicial restraint. 116 For instance, the Court has consistently applied a very narrow approach to what constitutes 'domestic remedies' by exempting applicants from exhausting all remedies other than those that are judicial in nature and that fall strictly within the ordinary domestic court structure. As such, it found that article 56(5) of the African Charter does not require applicants to utilise the processes of review and constitutional petition for the protection of fundamental rights -merely because the Court considers them to be extraordinary. 117 This approach is not consonant with judicial restraint as required by article 1 of the African Charter, 118 which allows states to exercise discretion as to the format and structure of the 'domestic remedies' they design to implement the Charter. The task of the Court should merely be to ensure that the 'domestic remedy' is available, effective and sufficient, and meets the standard of fairness in its operation. 119 The Court's case law is in line with this position, as illustrated in the APDH judgment on interpretation where it declined to guide Côte d'Ivoire on how to bring the composition of its electoral commission in line with its international obligations. 120 The Court arguably showed consistency in the case of Umuhoza v Rwanda where it undertook a wholesale factual and legal determination of which acts amount to a denial of genocide after domestic courts had ruled on the issue. 121 The problem is that in the APDH and Umuhoza cases the Court did not follow its own established precedent on deference and judicial restraint as set out in the earlier-cited Nguza judgment. A comparison between the Thomas, Mtikila, Abubakari, Umuhoza, APDH and Nguza judgments reveals inconsistencies in the Court's approach to deference and restraint.
Going by the Court's position on what constitutes a 'domestic remedy', most constitutional tribunals in Francophone or civil law Africa would fail to meet the standard of a valid remedy under article 56(5) as most of them do not belong to the ordinary judicial apparatus. 122 Yet, many of these institutions are vested with both material and personal jurisdiction to handle human rights complaints, including in the application of the African Charter. 123 Sadly, in years of law making, from the Mtikila judgment in 2013 to Thomas (2015), Abubakari (2016) or Owino (2017), the Court has not set clear and principle-based standards for assessing judicial and ordinary remedies.
A case in point on the nature of domestic remedies, particularly in civil law Africa, is that of Ajavon CRIET. Notably, in determining admissibility, the African Court set out the traditional two-prong test of an existing and effective remedy. Allegations of the applicant were in relation to administrative, criminal and constitutional remedies. Unfortunately, after establishing that the Constitutional Court existed as a remedy that the applicant could have approached, 124 the African Court omitted to assess the effectiveness of this remedy. In reaching the conclusion that the 'application cannot be dismissed for nonexhaustion of local remedies', the Court reasoned that the 'prospect of success of all existing remedies was marginal', 125 and 'inferred that the particular circumstances of the case rendered the remedies inaccessible and ineffective '. 126 Such reasoning and conclusion may be questionable in Benin's institutional context. The Constitutional Court of Benin has jurisdiction to adjudicate human rights petitions by individuals and groups. It has developed an extensive and rich case law including decisions made in direct enforcement of the African Charter. 127 In the context of the African Court's ruling on the merits in the Ajavon CRIET case, the Benin Constitutional Court in DCC 19-055 of 31 January 2019 declared unconstitutional article 12(2) of the law establishing the CRIET for being contrary to article 3 of the African Charter and article 26 of the Constitution. The Constitutional Court issued its ruling two months prior to the African Court's finding that the applicant in the Ajavon CRIET case did not need to approach the Constitutional Court as the 'circumstances of the case rendered the remedies inaccessible and ineffective'. 128 In its reasoning the African Court did not allude to or assess submissions by the respondent on the effective operation of the Constitutional Court. 129 Another admissibility-related law-making practice of the Court that does not sufficiently take into account state sovereignty and judicial deference is the unprincipled application of the theory of a 'bundle of rights'. The practice consists in declaring an application admissible on all issues raised by the applicant by bundling them together mainly on the ground that domestic courts ought to have been aware of other issues while examining only the one issue that was actually brought to their purview. 130 Arguably, there is a strong connection between claims of withdrawal for breach of sovereignty and an unprincipled observance of judicial restraint. An inconsistent or unprincipled application of article 1 of the African Charter might have caused the Court to assess facts and evidence previously adduced and determined before national courts in a jurisdictional fiction. 131 By doing so, the Court obviously was not cognisant of the 'fourth instance formula'. 132 Consequently, this practice may lead to an excessive or undue reliance on and interpretation of municipal law to ultimately establish a violation of international law. 133

Strategy-blind provisional orders
The practice of the African Court in relation to provisional orders raises issues. Recourse to provisional measures by its nature does not determine the merits of a matter. Orders in that regard should therefore be handled with extreme caution. The Court's practice in this respect is to assert jurisdiction prima facie, essentially in respect of material and personal jurisdiction. The Court, however, does not consider the prima facie admissibility of the matter. By at this initial stage determining the prima facie jurisdiction -but not the admissibility -of the matter, the Court's provisional measures order may override the admissibility of the matter in a way that causes unnecessary and unfair damage to the respondent. In the Kodeih case, for instance, it should from the onset have been obvious to the Court that remedies within the legal order of Benin still had to be exhausted and that the matter most likely was inadmissible. 134 In such an instance judicial economy and fairness demand a preliminarily ruling on admissibility. Doing so would have saved the Court the difficulty of having to, in its provisional measures order, first suspend the enforcement of a domestic decision with the costly associated consequences, only to eventually have to declare the matter inadmissible. It is curious why the Court embraces a quasi-systematic prima facie approach to jurisdiction but not to admissibility in instances such as this.
Another serious concern about the systematic or frequent use of provisional orders is that they may have such far-reaching impact that they supersede the impact of the eventual merits decision. For example, in the Ajavon Elections matter the burden incurred by the Court's order to suspend the election includes technical budgeting of US$12 million, 135 over two years of preparation, and the costs of campaigning. Should the alleged violations be established, the Court will issue further orders on the merits and reparation including monetary compensation and costs of potential legislative changes and rescheduling of the suspended elections. The Court's approach in this regard amounts to double jeopardy and is counter-productive in the framework of international human rights adjudication involving sovereign states.

Ruling by imperium versus substantiated reasoning
While the history of the establishment of the African human rights system commands a well-reasoned and purposive approach to judicial law making, 136 there seems to be more than just an impression that the Court often adjudicates based on imperium rather than by substantiated reasoning. In other words, a finding or ruling by the Court often is pronounced on the mere basis of the Court's say-so, rather than on a sound and thorough reasoning supporting a particular conclusion. This increasing trend may well have been a cause of concern to withdrawing states. 137 An illustration of this trend is the Court's inadequate reasoning in the Umuhoza ruling on the very issue of withdrawal. In its ruling the Court held that the declaration, and thus its withdrawal, constituted unilateral acts of the state outside the purview of the law of treaties. 138 However, the Court provided no reasons as to why and how a declaration should be considered a unilateral act in comparison with other traditional acts such as a declaration of war or the recognition of a state. Furthermore, the Court opted to read the words 'shall' and doivent as 'may' and peuvent, thus excluding possible political or judicial review of a state's refusal to make an article 34(6) declaration or to withdraw a declaration that it had filed. 139 The particular importance of this ruling required thorough and well-substantiated precedent-setting reasoning. The malaise ensuing from sloppy reasoning is illustrated by the Court's addendum to elaborate on the concerned holding, 140 to which Rwanda objected as one of its grounds of distrust in the independence of the Decisions of the Court in the Nguza, Abubakari and Penessis cases further illustrate a trend of ruling by imperium rather than by clearlyarticulated reasoning. Opinions issued by dissenting or concurring judges in the Ajavon CRIET judgments on the merits and reparations 142 raise the same concerns. 143

Reforming the African Court through structural and operational changes
In light of the ongoing assessment, there is a need to reform the Court to regain states' adherence, either through structural reform entailing re-designing some features of the court system, or by adapting elements of its practice. While some of the structure-related reforms should reasonably be contemplated in the mid-term, many operational and practice-related changes can be effected in the short term. These two categories of reforms are now briefly dealt with.

Redesigning a more balanced and purposive system
Systemic reform requires an understanding that the system reached early fatigue due mainly to imperfections in design. Institutional reform should aim at strengthening or adding benefits that balance the increased costs incurred by states as the system matures.

Structural changes: A two-tiered full-time Court
Ultimately, the AU should make the Court fully and permanently operational. The current part-time operation no doubt contributes to the rush in concluding deliberation and delivering decisions, challenges to adjudicate at the right time due to lower productivity versus constantly-rising dockets, substantive adjudication by orders, a dilemma between urgency and time-spread deliberations, adjudication fatigue due to repetitive deliberations, and slow and delayed justice owing to single-bench sittings. Making judges work full-time will also address the constraints of the quorum of seven Operating a permanent court cannot be done effectively without debating on the issue of recruitment of judges. For institutional harmony, it could be envisaged to adjust the nomination and election standards of judges on those applicable to elected officers of the AU Commission. The introduction of dossier selection and the interviewing of the candidate judges by former judges of the Court and other international recognised experts in the field would improve the management of issues of belongingness, productivity, administrative accountability, expedient and stable administration of justice, and perhaps also judicial law making. 145 An assessment also reveals that the cost of running a full-time African Court would be the same as, if not lower, than what the current format incurs at least with respect to the financial implication of judges' remuneration. 146 There also is a need to establish an appellate chamber in view of the Court's often-repeated position regarding the importance of the right to appeal in domestic proceedings. The need for an appeals chamber is also demonstrated by the increased number of applications and orders, the increase in the quantum of reparation orders, and the reduction in the overall time for implementation especially concerning states honouring dozens of orders. The rise of cases touching on 'critical' social or political issues also demands twotiered adjudication mechanisms, which largely applies expressly or by practice in other international regional or human rights courts. 147

Enhance state adherence and oversight
As alluded to earlier, it may be challenging to have the declaration repealed in light of responses to the various attempts so far made and in the current national and continental contexts. A more rewarding approach should be to manage the declaration regime by devising more incentives to join and to provide more alternatives to 144  withdrawing. New incentives such as the review, appeal, systematic amicable settlement and other tools certainly bring about a framework for greater adherence.
State involvement may also be improved by strengthening the framework for supervision and enforcement of the African Court's decisions. At the political level the system should be upgraded to a more effective process than the mere monitoring of the Court's report by the Executive Council under article 29(2) of the African Court Protocol. On the judicial front, article 31 of the Protocol should be transformed into an express competence for the Court to receive non-compliance applications. The framework document currently pending adoption by the Executive Council may be helpful in this respect as it provides approaches that include the judicial, political, diplomatic and technical monitoring of implementation. 148

Re-design complementarity with the African Commission
From a comparative perspective, the African Court Protocol assigns a critical role to the African Commission in a sui generis African regional system, compared to the European and Inter-American human rights regimes. While it may be true that the Protocol does not provide for a detailed complementarity clause, articles 2 and 8 leave adequate room for the Commission and Court to devise additional tools that best serve collaborative complementarity within the system.
A first tool could consist of including in the Protocol or Rules a non-limited list of matters for which the African Commission would be the corridor to the Court, such as grave and massive violations, electoral matters and amicable settlement procedures. Under this type of litigation the Commission could also serve as a shield or filter for the Court on some 'critical' cases. As a response to the issue of the Court reversing findings of the Commission when the latter files applications, it is a cause of satisfaction that the Commission's 2020 Rules of Procedure provide for the filing of cases immediately after seizure. 149 Conversely, the provision that the Commission must seek the authorisation of the complainant before referring the case to the Court may be regressive, bearing in mind cases of public interest or general law development that go beyond subjective rights. 150 The Commission could also consider filing cases after seizure when it has yet to make any substantive finding or by litigating the violation of article 1 of the African Charter in cases where states have not implemented its findings or requests for provisional measures.
Another tool is to unequivocally in its Rules allow original complainants at the African Commission full standing as 'coapplicants' before the Court. This approach holds the particular advantage of having civil society organisations relieve the Commission of the burden of costly litigation.

More contextual review and interpretation of judgments
Until its Protocol and Rules are amended, the African Court would gain by adjusting its practices on review and interpretation of judgments. By not objecting through amendments to years of jurisprudence, states have endorsed the interpretation of the law into opinio juris communis on issues such as the 'bundle of rights', 'extraordinary remedies' and 'prima facie jurisdiction'.
Based on this acceptance, the Court should be able to extend the 'review' authorisation in the Protocol. The Court should extend its competence to 'review … in the light of new evidence' not only to errors, whether material or factual, but also to omissions to rule, and to determinations not suitable in the circumstances of the case. The Protocol leaves room for the Court to frame Rules 67 of its 2010 Rules towards a much more contextualised and purposive review. In my view, it is within the Court's discretion to tailor the Protocol's flexibility into an innovative judicial law making that interprets 'new evidence' as being new facts, information, or even argument.
Judicial practice may also have to be adjusted with respect to the interpretation of judgments. For instance, the fact that the respondents requested for interpretation in the matters of Abubakari v Tanzania and APDH v Côte d'Ivoire should have been seen as a bona fide attempt to attain clarity. In the first case the Court interpreted its general order of remedying the violation as including the specific option of releasing the applicant, which it had declined to grant in the merits judgment. 151 Conversely, in the second case the Court declined to interpret the order of bringing the law in line with international norms on the ground that the request amounted to it indicating the manner in which to implement the merits judgment, which is the responsibility of the respondent. 152 The trend appears regressive and misguided. A state certainly would not seek interpretation of an order that is obviously vague if it did not intend implementing the decision. Therefore, the Court needs to carefully distinguish and consider interpretation requested in good faith in a manner that avoids reducing the likelihood of compliance and repeated adjudication on the same issues.

Devising practical alternatives to part-time operation
Inadequate judicial law making cannot be disconnected from rushed adjudication. The impression that the finalisation of matters does not receive sufficient attention cannot in turn be dissociated from the part-time operation of the Court. There should first be an inquiry into whether intersessional work at the Court has been as productive as it should be. 153 The higher number of decisions delivered from the time when the duration of sessions changed from two to four weeks cannot be an accurate parameter because the number of legal officers also almost doubled. Furthermore, most of the decisions accounting for the increase are provisional orders and procedural rulings that involve relatively or significantly less work than merit rulings. 154 Changing the picture requires innovative institutional adjustments that address constraints in the Protocol. The Court may need to adopt a situational interpretation to the part-time nature of its work. This approach could include setting up informal sections bound by the endorsement of the plenary, fully operating the judge-rapporteur mechanism, and enforcing the part-time scheme entirely by enhancing the work of judges during intersessions. The experience of remote operation and virtual sessions demanded by the COVID-19 lockdown has offered evidence that the Court can actually upgrade intersession work to its fullest and enhance its operation during inperson sessions. Doing so may demand a dramatic shift in working methods of both the registry and judges.

Purposive timing and mechanisms of adjudication
The African Court should improve its time management. It may need to adopt a more context and issue-based approach to the scheduling, deliberation and disposal of cases. It would, for instance, be more context-informed to not deliver in a row rulings on very sensitive matters touching on issues as policy-entrenched as nationality, the death penalty, and elections. This adjustment of course should be implemented on condition that deferring a ruling on a matter requiring changes to the legislation on the death penalty but not the execution of that sentence would cause no or marginal impact.
With respect to electoral matters in particular, adjudication demands expediency. This applies to the use of suo motu powers bearing in mind competing international obligations of the states involved and domestic constraints. An effective time setting may require a thorough balancing of interests involved and drawing of the state's discretion or margin of action in implementing the African Charter. In this instance, adjudicating public international law in an era of increased sovereignty must give due consideration to topics relating to governance, such as elections or political disputes, which come before the Court under the cover of human rights. The proposal is not that the Court should compromise on fairness and justice, but it cannot be blind to the internal operation of states while adjudicating matters brought before it. 155 Urgency, expediency, special circumstances and fairness will remain the exceptions.
While dealing with cases that have a bearing on states' discretion, one recipe could also be to use more contextualised mechanisms such as water-testing adjudication, incremental law making, and purposive jurisdiction or admissibility filtering. These tools should be devised and used in a manner that takes proper cognisance of the domestic system, avoiding giving the impression that decisions of the Court are materially impossible to implement in the domestic context or render useless the remedy afforded to the applicants. In the same vein, the Court should embrace both effective task segregation between itself and domestic actors and adopt passive 155 Soubeyrol (n 12) 239. virtues. 156 Well-known practices include judicial deference and prudential self-restraint. 157 Similarly, there is a need to assess the administrative prima facie probe of admissibility by the registry and ensure that it informs options taken by the Court while examining requests for provisional measures. The purpose is to embrace a more cautious exercise of prima facie jurisdiction on a case-by-case basis.
In order to alleviate the litigation burden of respondent states, a change in practice may also gain from an enhanced recourse to more protracted and systematic amicable settlements, the merger of cases, and pilot judgment procedures. These methods may be more relevant where the cases involved are highly political in nature, may widely affect uninterrupted governance, or involve serious and massive violations.

Strategic selection of most appropriate remedy
In adjudication it may be critical to choose the most appropriate remedy according to timing, circumstances and domestic context. In the case of the African Court, there may not be a need to expressly include strategic remedy selection in its statutes, but rather to frame them in internal practice guidelines that would evolve over time and according to different circumstances. Such directive, for instance, would serve to guide the adoption of provisional orders in a manner that avoids any clashes with principles of fairness, equity, equality of arms, and balancing of all relevant constraints, interests and rights involved.
The African Court should also avoid making orders ultra petita as they may extend beyond what is necessary to safeguard the interests of the applicant, or may have the effect of prejudging the merits of the matter. The role of a fair adjudicator is not to grant a remedy merely because the applicant asked for it. Mostly, a range of possible remedies is open to an adjudicator in solving a dispute. The challenge is to choose from among these available solutions the most suitable remedy under the particular circumstances. Hypothetically, in the Ajavon Local Elections case a more sober adjudicator would have borne in mind the emergency of an electoral matter, shortened the time of adjudication to fewer days, resorted to its suo motu powers to issue orders, and done so well ahead of time. The situation could have been totally different had the Court, for example, issued its order 90 days instead of 30 days before the elections. Substantively, the Court should have also limited the order to the applicant and potentially his political party, instead of bringing a whole country to a standstill. Finally, the order subsequently issued by the Court in the Houngue case is evidence that the Court did not make the most suitable order in the Ajavon Local Elections case. In the Houngue case the Court ordered the same respondent to take measures to ensure the applicant's participation in the very elections it had just suspended.
The same argument applies with even greater force to the order for payment of US$ 66 million issued in the Ajavon CRIET case. 158 The African Court should have been context-conscious that, as a matter of practice, states' directorates of litigation are budget-set by legislatures with a specific ceiling in terms of the amount of damages to disburse in a given year. 159 The Court is putting its legitimacy at stake if it gives the impression that it makes orders that are not likely to be implemented. Numerous orders due by Tanzania, including multiple monetary injunctions, also fall into this category, bearing in mind that Tanzania has implemented fewer than 15 per cent of the 60 orders so far issued by the Court against it. 160

Judicial diplomacy
Judicial diplomacy cannot be overlooked as a critical ingredient of the effective operation of an international human rights tribunal. Bearing in mind that international human rights litigation is a relatively recent phenomenon in Africa, the bureau and registry of the Court should on their own motion or jointly with the AU policy organs set out a framework for regular engagement with states. In its formation and effective implementation, international human rights law and its enforcement mechanisms undeniably are state-led. In contrast to their domestic counterparts, international human rights tribunals are sui generis in their establishment, the law that they enforce, and the litigants before them. A sign that judicial diplomacy remains a workable option even within the current context of crisis is that none of the four withdrawing states denounced the African Court Protocol. In fact, Benin and Côte d'Ivoire expressly stated that they remain parties to the Protocol, while Benin announced proposals for reform. Upon withdrawal of their article 34(6) declarations, Rwanda and Tanzania indicated that they intended to file the declaration afresh after a process of review. 161 Judicial diplomacy should extend on a more systematic basis to civil society, including the academia and media, with the aim of regularly assessing the work of the Court and devising preventive means to sustain state adherence. Ultimately, the Court should set a warning and response system to allow for the airing of contestation. It does not appear that such has been done between the first withdrawal in 2016, and the three that occurred four years later within the span of six months. A practical way of implementing this proposal would be to establish an assembly of state parties as is known under the International Criminal Court regime.

Conclusion
It may be contended that Rwanda, Tanzania, Benin and Côte d'Ivoire have withdrawn from the African Court's direct individual jurisdiction to escape accountability. 162 However, this article provides ample evidence that these states' disengagement was caused by factors that go beyond the specifics of the cases against them, and extend to challenges inherent to the system, and to the practice of the Court. The issues raised call for appropriate adjustments, as proposed and discussed above.
In terms of process, the proposed reforms may be designed and implemented within the framework of the larger reform process led by the AU, or as part of a separate process initiated by the Court. As a matter of strategy, it is in my view advisable that the Court takes the lead -as it is entitled to do under article 35(2) of the Protocol -so as to retain oversight both in terms of process and substance. Should states take the lead, they may legitimately tend to frame and effect changes informed by their misgivings, which may not offer the guarantee of a technically fair, objective and system oriented reform process.
Although this did not form part of the discussion, the importance of civil society in supporting cooperation between the AU, states and the African Court cannot be overemphasised. Civil society organisations also have a role in preserving the Court as a public service institution through accountability checks and by advocating reform that is likely to reinforce states' full adherence to the Court's jurisdiction.
The current political context in Africa suggests that if the Court is to survive as a source of legitimate supervisory authority, its support system must ensure a sustainable balancing of the various interests involved. Judicial law making should therefore evolve while balancing judicial activism and restraint. This approach requires going by the pace of democratic and constitutional governance on the continent. Going by the current state of affairs, there is no indication that the African Court of Justice and Human and Peoples' Rights will operate any time soon. 163 In any event, its provision granting immunity to incumbent heads of state and other challenges cast some doubt upon its future functioning. Hence, the most reasonable option is to improve the design and practice of the African Court, the only continental justice system that translates the AU's Agenda 2063 of a continent where justice, human rights and the rule of law matter. Both structural changes and adjustments to judicial practice are required to save the remaining article 34(6) declarations, to get the withdrawing states to return to the fold, to secure further adherence, and to restore the African Court's weakened legitimacy.