In late August 2017, the Supreme Court of Kenya heard a challenge to the presidential elections held just days earlier.Footnote 1 In a remarkable result, it affirmed the petitioners’ complaints of electoral irregularity and nullified the entire election, ordering a fresh election within 60 days. While such a degree of judicial intervention into the electoral process is rare and surprising on its own, this decision was especially stunning because just four years earlier the Supreme Court had dismissed a petition alleging similar electoral irregularities with nary a glance at the merits of the case.
In this paper we ask two questions. First, why did the Supreme Court respond so differently in these two cases? We argue that much of the explanation lies in the changing levels of public trust in the two institutions involved: the courts on one hand and the Independent Electoral and Boundaries Commission (IEBC) on the other.
As an institution established by the 2010 Constitution, in 2013 the IEBC still basked in the afterglow of Kenya's successful constitutional transition. Public trust in the IEBC was high. The judiciary, on the other hand, had only just begun a process of transformation after decades of executive influence and partisanship.Footnote 2 A precariously positioned Supreme Court was not confident to stick its neck out against a trusted IEBC. By 2017, however, the process of judicial vetting had restored public trust in the courts, while the IEBC had become mired in corruption scandals and perceptions of partisanship.Footnote 3
Second, we use the Kenyan case study to ask when it might be appropriate, more broadly, for courts to scrutinise the merits of an independent electoral management board's decisions, or indeed the decisions of any of the independent institutions that safeguard constitutional democracy. The entrenchment of an electoral management board (EMB) in Kenya's 2010 Constitution is consistent with a broader move in recent years to establish a ‘fourth branch’ of ‘institutions protecting democracy’ (or IPDs).Footnote 4
The addition of a fourth branch to a centuries-old model of constitutional democracy, however, complicates the web of relationships between those branches. In this paper we focus on one dyad in this web: the relationship between an IPD and an apex court. We sketch the outlines of a new model of judicial intervention based on a distinction we draw between the ‘constitutive’ and ‘regulative’ functions that EMBs and other IPDs fulfil.Footnote 5
We argue that when an IPD acts in ways that constitute rights – that is, when an IPD makes it possible for people to exercise constitutional rights – judicial intervention is justified. Democratic rights to vote and stand in elections, for example, are rights that require constitutive conduct because they cannot be exercised without the machinery of an electoral system in place, including institutions mandated to operate that machinery. On the other hand, when an IPD acts in a regulative manner, for example by safeguarding already constituted rights against infringement from other state actors, courts should be less inclined to intervene to set aside IPD decisions.
Our model is prescriptive. It suggests when courts should intervene and when they should defer to an IPD, but it does not claim to explain why courts have acted or how they will act. Certainly, our model does not explain the Kenyan Supreme Court's vastly different responses to the 2013 and 2017 presidential elections. The IEBC fulfilled a constitutive function in both situations, and we think that the Court should have approached both petitions with a similar degree of scrutiny, even if ultimately to uphold the 2013 election.
As a prescriptive account of the relationship between courts and EMBs, then, our model runs up against the realities of pragmatic realpolitik. In Kenya, in the early 2010s, the balance of public trust between the IEBC and the judiciary explains the Court's reluctance to intervene in the 2013 election. The fact that the 2013 election was concluded without significant electoral violence – a far cry from 2007's post-election violence – should not be discounted either. The success of the 2013 election was always going to be measured by the absence of violence.
As another aspect of realpolitik, it is often easier for courts to intervene in low-stakes regulative IPD decisions than in the high profile, marquee actions that are constitutive of constitutional rights. This too runs counter to how we propose courts should act. Ultimately, ours is an argument of constitutional principle. We offer an aspirational, theoretically grounded model for structuring the relationship between courts and other institutions protecting constitutional democracy, but we recognise that our model may not be reflected in the choices that courts actually make.
Finally by way of introduction, it's worth noting that the Court's intervention in 2017 was completely ineffective. The successful petitioner, Raila Odinga, ultimately refused to participate in the second election citing concerns about the IEBC's integrity. His withdrawal meant that his rival, Uhuru Kenyatta, would retain the presidency. The irony is that, having found the IEBC to have mishandled the election in the first place, the Court remitted the election back to the same compromised institution for the rerun. Although a stunning judicial result, the real-world outcome of the 2017 decision was a disappointing reminder that constitutional principle does not always triumph over constitutional practicality.
Outline
In Part II of the paper, we consider the development of IPDs and the fourth branch of government as a feature of modern constitutional democracy. We examine how a distinction between IPDs established by ordinary legislation on one hand and by the constitution on the other, frames a common understanding of their relationship with courts that borrows heavily from orthodox administrative law, and its distinction between appeal and review.
Overlaying this administrative law framework on the fourth branch of government, however, has an ambiguous effect. It does not give a clear answer as to when courts should apply intensive appellate scrutiny to IPD decisions, and when they should show administrative law deference. Indeed, two leading scholars in the field, Mark Tushnet and Michael Pal, adopt apparently different positions in this regard. In Part III, we try to resolve the ambiguity by suggesting that it is not the constitutional or statutory foundation of an IPD that matters, but whether its functions are constitutive or regulative.
Finally in Part IV, we use our functional model to evaluate the Kenyan judiciary's responses to IEBC decisions. As a matter of institutional design, the IEBC has both constitutive and regulative functions, and while some court decisions are consistent with our model's prescriptions, others – like the Supreme Court's 2013 decision – are not. The point of this evaluation is not to criticise the Kenyan Supreme Court, but to emphasise that courts are sometimes swayed by pragmatic considerations even though constitutional principle might pull in another direction.
Framing a new relationship between courts and IPDs
The emergence of the fourth branch of government
Government oversight is not a new idea. Indeed, the very objective of separating the functions of government between three branches is to ensure that each can act as a counterweight to the power of the others and limit their conduct – hence the phrase ‘checks and balances’. In both parliamentary and presidential systems, legislatures scrutinise executive conduct and can remove members of the executive branch in cases of extreme wrongdoing. In most common law legal systems, courts are empowered to assess the lawfulness of executive and administrative action.
The experience of the late twentieth century and the start of the twenty-first, however, has suggested that the system may not work as well as we might have hoped. Although Francis Fukuyama proclaimed the unassailable victory of liberal democracy in the early 1990s,Footnote 6 the more recent phenomena of ‘illiberal democracy’,Footnote 7 ‘competitive authoritarianism’Footnote 8 and ‘democratic backsliding’Footnote 9 suggests that the traditional model of divided government is less effective than we might have hoped at safeguarding the core principles of constitutional democracy. Ad-hoc or standing committees of the legislature, parliamentary question time, and courts more commonly asked to resolve civil disputes and enforce criminal law, have all struggled to rein in leaders bent on extending or consolidating executive power.
One response has been to supplement these traditional mechanisms of legislative and judicial oversight with a brand new kind of democracy-guarding institution. Constitutional courts staffed by specialist jurists and with specific jurisdiction have been established for the purpose of ‘defending the constitution itself’, for example.Footnote 10 The widespread acceptance of constitutional courts opened the door to the idea that a specialist institution could be dedicated to protecting constitutional rules and principles. Agencies tasked with rooting out corruption, ensuring accountability in government spending, and investigating rights abuses at the hands of government emerged throughout the twentieth century – what we now call IPDs.Footnote 11
Establishing these democracy-protecting agencies in ordinary legislation has its flaws, however. It is all too easy for elected legislators to pass laws that increase their chances of re-election, by drawing electoral districts in ways that favour them,Footnote 12 or making it more difficult for people to vote.Footnote 13 Experience has shown that nominally independent elections management bodies established by statute are just as open to self-interested manipulation. Michael Pal suggests that when EMBs are ‘born of regular legislation that defines their existence, functions, authority, and appointments process’, they are vulnerable to partisan capture by political majorities.Footnote 14 Under section 41 of Kenya's easily and frequently amended 1963 Constitution, for example, the president was empowered to appoint and replace all the commissioners of the IEBC's predecessor, the Electoral Commission of Kenya (ECK).Footnote 15 The ECK was widely seen as a co-opted and partisan institution.Footnote 16
An important development in the architecture of IPDs was South Africa's 1996 Constitution, which entrenched a handful of these integrity institutions in a chapter entitled ‘State Institutions Supporting Constitutional Democracy’.Footnote 17 The decision to root an IPD in an entrenched constitution with a relatively difficult amendment rule, rather than in legislation, executive order or an easily amendable constitution, lends it a degree of permanence and independence. We might in turn expect such IPDs to be more effective in keeping the elected branches of government within constitutional limits.Footnote 18 Whether this turns out to be true, of course, is an empirical question that we do not propose to answer here. It does seem fair to conclude, though, that recognising the important role IPDs play in constitutional democracies has led to an increasing tendency to entrench them in the constitution rather than merely in legislation.
Our focus here is on how the relationship between the courts and IPDs is influenced by institutional design. More than this, we suggest that whether an IPD has constitutional or statutory foundations does not provide a useful guide to framing its relationship with the courts. Rather, what matters is whether an IPD fulfils a constitutive or a regulative function. We expound on this novel distinction in Part III below, but first, we consider how the distinction between constitutional and statutory IPDs invites a rough overlay of orthodox administrative law principles onto the IPD-court relationship.
An (ambiguous) administrative law model for inter-branch relations
In common law systems, the courts have always retained jurisdiction to review the decisions of administrative agencies for compliance with the principles of procedural fairness (or natural justice) and to ensure they do not act beyond the limits of their empowering statutes. This is the doctrine of ultra vires, or the principle of lawfulness.
Whether courts should be engaged in reviewing administrative decisions on the substantive merits, however, is a more vexed question. For courts of review to reconsider the substance of administrative decisions, in the way that courts of appeal reconsider the judgments of the courts below them, might allow review courts to usurp the decision-making power that the legislature has conferred on the administrative branch and undermine the very separation of powers that constitutional democracy requires. The exercise of something approaching appellate jurisdiction, or what is referred to as ‘correctness’ review in commonwealth jurisdictions, is usually confined to very limited circumstances.
Distinguishing between appeal and review in this way is meant to uphold democracy on one hand and the rule of law on the other. Close judicial scrutiny over whether administrative decisions are within the limits of empowering statutory provisions upholds the rule of law, while shielding administrative decisions from close scrutiny of their substantive merits upholds a democratically representative legislature's decision to confer decision-making power on specific agents.
As David Dyzenhaus has long argued, however, this dichotomy between the rule of law and democracy, and in turn the distinction between different standards of judicial scrutiny, is conceptually difficult to maintain.Footnote 19 Indeed, the entwinement of democracy and the rule of law is part of the very fabric of constitutional democracy. Consider that any statute passed by a democratically elected and representative legislature carries the imprimatur of the people's democratic will, but that in a constitutional democracy those statutes must be congruent with the provisions and principles of the constitution. Just as administrative decisionmakers are bound by the statutes that empower them, a legislature (and the democratic will it represents) is bound by basic constitutional rules and principles.
As a result, for courts to keep administrative action within the limits of a democratically enacted empowering statute is an exercise in both democracy and the rule of law. To suggest that courts uphold either democracy or the rule of law when applying one or some other standard of scrutiny, elides the close connection between the two principles. The 1898 decision in Kruse v Johnson recognises the connection between substantive review and lawfulness review, holding that Parliament cannot have intended to confer a power on an administrative decisionmaker to act unreasonably.Footnote 20 An unreasonable decision is necessarily unlawful.
In many common law jurisdictions the story of expanding reasonableness review is similar.Footnote 21 To the extent that IPDs in general and EMBs in particular are mere creatures of statute and occupy a position in the constellation of administrative agencies alongside town councils, film review boards and pharmaceutical safety commissions, we might expect their decisions to be subject to the same degree of judicial scrutiny on the grounds of procedural fairness, lawfulness and substantive reasonableness.
The administrative law approach runs through some of Mark Tushnet's recent and important work on the relationship between the emergent fourth branch and the courts. He suggests that constitutional courts will ‘routinely exercise supervisory power’ over IPDs, and that they ‘will never relinquish complete control over the operation of other IPDs.’ If IPDs are to remain meaningful in the face of judicial oversight of the substantive merit of their decisions, Tushnet finds that ‘the best solution…is for constitutional courts to give substantial deference to jurisdictional and procedural choices made by IPDs’.Footnote 22 Indeed, this is largely how the US Supreme Court has attempted to resolve the question of judicial intervention in administrative decisionmaking.Footnote 23
How courts should review other government actors’ decisions is a more general and longstanding problem in administrative law. As is commonly the case in the administrative branch, the work of IPDs often involves expertise that is not reflected in the judiciary.Footnote 24 This reveals the administrative law model's ambiguity: on one hand, courts should defer to substantive IPD decisions because they lack the expertise and experience IPDs have; on the other hand, substantively unreasonable IPD decisions should be just as open to judicial review as any other unreasonable administrative decision.
At the same time, the entrenchment of IPDs in the constitution beckons the resurrection of a stark distinction between appeal and review. Michael Pal argues that if IPDs comprise a fourth branch of government equivalent in constitutional status to the judiciary, legislature and executive, and which fits into the systems of checks and balances along with them, then IPDs should not be held directly accountable to any of the other three branches of government.Footnote 25 Indeed many of the constitutional provisions establishing EMBs expressly insulate them from judicial oversight.Footnote 26 On this understanding, the decisions of the fourth branch of government deserve substantial deference from the courts. Courts should not sit in appeal over IPDs and substitute their own views for those of the IPD.
This difference in how two leading scholars apply the same set of administrative law ideas suggests that the administrative law model, by itself, is unlikely to provide a useful basis for framing a relationship between IPDs and courts. One way out of this dilemma is to focus on the function that an IPD fulfils in a constitutional democracy, particularly whether it facilitates the exercise of rights or protects them from infringement, rather than whether it has statutory or constitutional origins.
Constitutive and regulative functions
Constituting democratic rights, regulating liberty rights
In The Concept of Law, Herbert Hart describes how a legal system constitutes the categories of lawgiver and legal subject even as it regulates conduct in society. People in a society rely on a set of secondary ‘rules of rulemaking’ to discern which primary, conduct-governing rules count as law that must be obeyed. If a government or sovereign expects its subjects to know which rules to obey, it must follow these secondary rules when it purports to make law. A society's secondary rules ‘are constitutive of the sovereign’.Footnote 27
Primary legal rules, for their part, can be regulative or constitutive in narrower ways. A town bylaw might prohibit riding a bicycle on the pavement (or sidewalk),Footnote 28 but the bylaw is not constitutive of bicycle-riding. You can ride a bike whether or not the bylaw exists. Rather, the bylaw regulates how people may legally do something they are otherwise free to do.
For an example of a primary rule that has a constitutive function, think of electoral law. Voting is not something you can do in the absence of an electoral system and all the legal and institutional machinery that establishes it. You can vote only once an electoral system is in place, with rules that tell you what the various requirements and practicalities of voting are. Electoral law itself is constitutive of the very possibility of voting.Footnote 29
In most constitutional democracies, moreover, electoral law elaborates on a set of basic constitutional rights to participate in the public life of the community: rights to vote in free, fair and regular elections, stand for public office, engage in political speech, and form groups or parties for the purposes of advancing policy objectives.Footnote 30 Electoral law takes its cue from the constitutionalisation of democratic rights, and its content is closely informed by those rights. Indeed, where electoral law departs from or fails to fully uphold constitutional democratic rights, courts have declared them to be unconstitutional and invalid.Footnote 31
The electoral system, built up by electoral law and reflecting constitutional rights to democratic participation, is constitutive of these democratic rights. In turn, the EMB that establishes and manages an electoral system performs a constitutive function. Michael Pal puts it as follows:
The fourth branch model notably recognizes that institutions are required to breathe life into rights, including democratic ones. It moves beyond protecting certain electoral practices to guaranteeing a particular institutional setup for election administration. Rights to democratic participation – such as the right to vote, stand as a candidate, or engage in political speech – are insufficient on their own in this model.Footnote 32
Democratic rights are different from other constitutional rights precisely because they require a body of electoral laws and the machinery of an electoral system in order for people to exercise them. In contrast, the rights that most other IPDs are tasked with upholding are already held by people and enforceable against the world. Rights not to be unfairly discriminated against, to security and liberty, to fair labour practices and to practice a religion freely all guarantee a sphere of personal freedom insulated against governmental (and sometimes private) interference. These rights delimit the state's authority over its subjects, and IPDs protect these rights by regulating the state's exercise of its authority.Footnote 33
The rights at stake in disputes before human rights tribunals, auditors general, anti-corruption agencies and police oversight units remain meaningful and enforceable even without these IPDs. Unlike democratic rights, they do not depend on a special framework (like an electoral system) or an institution's functions (like an EMB running an election) in order for people to enjoy them. Protecting and enforcing these rights requires an IPD to act in a regulative capacity, rather than a constitutive one.
Although we have so far suggested a stark line between constitutive and regulative functions, it may be more accurate to think of IPDs as performing mostly regulative or mostly constitutive functions. A single IPD decision may contain elements that are both regulative and constitutive, and fall somewhere on a spectrum between purely constitutive and purely regulative. For example, the resolution of a dispute about the application of voter ID laws both regulates elections officials’ application of voter laws while helping to constitute the category of legal voter.
Whether the distinction is conceived of as spectral or binary, it nevertheless provides the foundation of the model of judicial engagement that we propose below. At the same time, the distinction offers a nuanced and comprehensible way to reconcile the difference in approach between Tushnet and Pal.
A new model for judicial engagement
Recall that Tushnet takes the position that courts should retain oversight of the substantive merits of IPD decisions, while deferring to decisions of a procedural nature. Pal on the other hand reasons that courts should defer to the substantive decisions that constitutionally entrenched ‘fourth branch’ IPDs make, and intervene only in procedural matters. When IPDs have merely statutory foundation, though, Pal concludes courts should engage in substantive review as they do with any other administrative agency.
Our suggestion is that the distinction that matters is not the constitutional or statutory origin of an IPD, or even the substantive or procedural character of a decision, but whether an IPD's decision is constitutive or regulative of rights. When an IPD acts in a constitutive capacity, courts should retain oversight of the substantive merits of that IPD's decisions in the way that Tushnet suggests, exercising a degree of scrutiny similar to an appellate court. But when an IPD acts in a regulative capacity, courts should review the decision as it would any other administrative decision, on the familiar administrative law grounds of procedural fairness, lawfulness and reasonableness.
Our justification for this new approach is on one hand ontological, and on the other hand an extension of Pal's equivalency thesis. From an ontological perspective, EMBs are usually tasked with both running elections and certifying their integrity. To refuse to allow a court to review the substance of an EMB's declaration that an election was fair leaves the EMB as the only guarantor of its own competence. It stands somewhat to reason that if the EMB was incapable of running a fair election in the first place, we might want to be sceptical about its post-facto declaration that the election was fair. The ontological problem is that we can never know anything about the fairness of the election – resting as it does on the competence of the EMB – if all we have is the EMB's own declaration that it is competent and that the election was fair.Footnote 34
The self-referential, bootstrappy nature of this arrangement is a worry precisely because the primary actor on which our democratic rights depend is the EMB. We cannot exercise our democratic rights except through the electoral machinery that the EMB operates. When an EMB makes an error that undermines these democratic rights and courts are unavailable to scrutinise the EMB's conduct, voters will be left with no forum in which to vindicate their rights.
When some state actor performs the primary act that threatens constitutional rights, an IPD may be called on to perform a regulative function (in some cases IPDs may even have jurisdiction over rights-limiting conduct in the private sphere). A human rights commission is asked to intervene, for instance, only after a person is alleged to have discriminated against a certain group. A police inspector general is asked to investigate only after police are alleged to have improperly executed a warrant. An anti-corruption agency acts only when public funds may have been misappropriated, and so on. These IPDs all play secondary review roles, scrutinising a primary act by a different branch of government that affects rights that people already hold and enjoy.
This observation captures our argument's institutional dimension. To ensure that constitutional rights are respected and protected in a government structure that separates power, the relationship between the branches of government should be such that there is at least one forum of oversight for each rights-affecting act. Where the police execute a warrant improperly or government officials commit corruption, fourth-branch institutions exist to protect the relevant affected rights. For voting rights, which must be constituted by an electoral system and some institutional actor in the first place, the fourth branch tends not to play an oversight role. It is true that in some situations the EMB might not be the institution the actions of which constitute democratic rights: the executive branch may play a role in elections or intervene in a specific election, and the EMB may in these cases exercise some degree of oversight. But where the EMB's conduct constitutes rights, our concern is that there is rarely another fourth-branch institution capable of scrutinising the EMB's conduct.
We can also put the point in terms of Pal's equivalency thesis. If we take seriously the idea that fourth-branch IPDs occupy a position similar in status to the other three branches of government, then it follows that all four branches should be subject to a similar degree of substantive oversight. Along with the courts, IPDs other than EMBs scrutinise the conduct of the executive, legislative and administrative branches when it affects constitutional rights. When an EMB's conduct affects democratic rights, equivalency demands that some other branch of government be empowered to scrutinise those decisions. The judiciary already plays an oversight role with respect to the legislature's rights-affecting decisions, and there seems to be no reason it should not scrutinise an EMB's rights-affecting conduct.
We end this part of the paper with a point we take up in our examination of Kenya's experience: EMBs perform constitutive as well as regulative functions. When an EMB acts as a tribunal for the resolution of electoral disputes, for example, the EMB is itself a point of secondary oversight that regulates primary conduct affecting democratic rights. On our normative model, there is less justification for substantive judicial intervention here because the EMB already provides a point of secondary oversight. Courts should be prepared to let EMBs get on with these regulative functions, overseeing their decisions with something more akin to administrative law review than strict appellate scrutiny.
In any particular arrangement, however, questions will remain about how well insulated a particular EMB's dispute-resolution branch is from its elections-management branch. Kenya's IEBC, for example, has established two distinct dispute resolution tribunals, the Electoral Code of Conduct Enforcement Committee and the Dispute Resolution Committee. Although staffed by IEBC commissioners, both bodies are constituted differently from the IEBC itself and are intended to operate with some degree of institutional independence. The general point about this separation of functions, however it may happen, is that courts may not need to subject an IPD's regulative functions to searching substantive scrutiny if the fourth branch is already able to exercise at least some degree of meaningful secondary scrutiny. We use the relationship between Kenya's courts and its IEBC in the 2010s to explore this dynamic.
Evaluating electoral reticence and activism in Kenya's judiciary
Kenya's 2010 Constitution provides that the IEBCs is responsible for ‘conducting or supervising referenda and elections to any elective body or office established by the Constitution’ (Article 88(4)). Further, at every election, the IEBC must ensure that the voting system is ‘simple, accurate, verifiable, secure, accountable and transparent’ (Article 86(a)), that the electoral system delivers ‘free and fair elections’ and is ‘administered in an impartial, neutral, efficient, accurate and accountable manner’ (Article 81(e)). In order to fulfil this mandate, Article 88(4) lists a number of more specific IEBC functions:
a. the continuous registration of citizens as voters;
b. the regular revision of the voters' roll;
c. the delimitation of constituencies and wards;
d. the regulation of the process by which parties nominate candidates for elections;
e. the settlement of electoral disputes, including disputes relating to or arising from nominations but excluding election petitions and disputes subsequent to the declaration of election results;
f. the registration of candidates for election;
g. voter education;
h. the facilitation of the observation, monitoring and evaluation of elections;
i. the regulation of the amount of money that may be spent by or on behalf of a candidate or party in respect of any election;
j. the development of a code of conduct for candidates and parties contesting elections; and
k. the monitoring of compliance with the legislation required by Article 82(1)(b) relating to nomination of candidates by parties.
The performance of some but not necessarily all of these functions is constitutive of democratic rights. Without the IEBC impartially and efficiently conducting simple, transparent and accurate elections, registering voters and candidates, delimiting electoral boundaries, or observing and monitoring the elections, it is not clear how rights to vote could, practically speaking, be exercised.
Certain of these functions are more regulative in nature, however. They require the IEBC to consider how the exercise of democratic rights may have been affected at other points in the electoral system, rather than to construct or operate the machinery for exercising those rights in the first place. The standout example of the IEBC's regulative functions is paragraph (e) above: the settlement of electoral disputes (other than those challenging the declaration of results). The operative distinction lies in the fact that when the IEBC is called on to settle a dispute, for example going to the nomination of candidates or some other intermediary step in the electoral process, the parties approaching the IEBC are alleging that their already-constituted democratic rights are under threat. In these cases, the IEBC's dispute tribunal acts to regulate the conduct of another actor in the electoral system.
The regulation of the process by which parties nominate candidates for election (paragraph (d)) is also more regulative than constitutive, because the primary rights-affecting conduct – selecting candidates for election – happens within the walls of political parties themselves. In this case democratic rights are primarily affected by the internal party process rather than the IEBC's conduct. On the other hand, where an EMB (Kenya's IEBC or some other EMB) makes rules that govern how parties can nominate candidates, the EMB's rules would tend towards the constitutive rather than the regulative end of the spectrum.
Seeing each of the IEBC's functions as predominantly constitutive or regulative helps us to evaluate the Kenyan Supreme Court's justifications for judicial intervention or deference in electoral challenges. As a normative point, we do not think it is justifiable for the Supreme Court to decline to engage in deep and critical scrutiny of the IEBC's constitutive functions – although there may be compelling reasons of political pragmatism to refrain from doing so. More broadly, we suggest that the relationship between EMBs (or other IPDs) and courts should be structured with this functional distinction in mind. EMBs should expect to have the courts take a close look at their constitutive functions (or in terms of judicial procedure, subject them to appeal on the merits), while being left to perform their regulative functions subject only to administrative law review.
Judicial reticence: the 2013 decision and review of IEBC tribunal decisions
The Kenyan Supreme Court's responses to the petitions challenging the presidential election results in 2013 and 2017 are at polar extremes, even though the facts of both cases are remarkably similar. Uhuru Kenyatta and Raila Odinga were the leading presidential candidates both times, Kenyatta appeared to have won by a whisker both times, and Odinga petitioned the Court both times. In 2013, the Court had little to say beyond confirming the IEBC's declaration of Kenyatta as the winner, but in 2017 the Court interrogated the evidence and set aside the entire election. We look first at the Court's restrained response in 2013.
The 2013 election, Kenya's first under the 2010 Constitution, took place on 4 March 2013. There was a record turnout, which saw over 12 million Kenyans cast their votes.Footnote 35 After the newly formed IEBC declared that Uhuru Kenyatta had won, passing the threshold to avoid a runoff by only 8400 votes, Raila Odinga petitioned the Supreme Court with allegations of irregularities.Footnote 36
The Court rejected Odinga's challenge and upheld the IEBC's declaration of the results.Footnote 37 In stark contrast to the 2017 judgment, the Court was highly deferential to the IEBC. Indeed, the need for deference was its primary justification for dismissing Odinga's petition, even though the Court did not justify its commitment to deference in the first place. On the contrary, the Court appeared merely to have accepted the IEBC's submission that courts owe the IEBC deference. In a section of the judgment headed ‘Judicial restraint’, the Court paraphrased counsel for the IEBC's arguments:
Learned counsel…has called for the adoption of restraint by the Court, in this Presidential election matter. He urges that the facts and special circumstances of this case require restraint, in the judicial approach.Footnote 38
The Court then accepted two lines of argument urged by the IEBC. First, it agreed that the elections must have been free and fair simply because, unlike 2007 and happily so, there had been no post-election violence. Second, the Court accepted the IEBC's position that courts should in general refuse to intervene in political matters. Relying on the US Supreme Court's decision in Bush v Gore and the South African Constitutional Court's decision in Minister of Health v Treatment Action Campaign, the Court concluded that it should show restraint in order to nurture public trust in the judiciary and promote national stability, especially in a new and fragile democracy.Footnote 39 Instead of intervening in the presidential election, the Court held, it is better for the Court to preserve its ‘political capital’ until it really needs it.Footnote 40
The Court seemed eager to find additional reasons to avoid the merits of the petitioners’ case. The burden of proof the Court imposed on the petitioners, for example, is ‘almost insuperable’Footnote 41 and at odds with established jurisprudence and procedural rules about standards of proof. Similarly, the Court preferred inflexible interpretations of the time limits and rules for the admissibility of evidence, despite the injunction in Article 159(2)(d) of the Constitution that ‘justice shall be administered without undue regard to procedural technicalities’.Footnote 42
When the Court did finally get to the petitioners’ core substantive complaint, its reasoning for rejecting it was thin. The petitioners’ argument was that the constitutional requirement that a presidential candidate win ‘more than half of all the votes cast in the election’ should be understood to include spoiled ballots.Footnote 43 The Court disagreed, however, concluding that spoiled ballots should not be counted in the calculation of whether a candidate has passed the 50 per cent threshold necessary to avoid a runoff.Footnote 44
The Court did not offer much in the way of reasoning to support its interpretation, despite having been asked by the petitioners to provide a ‘guiding interpretation’. Instead, the Court stated that in taking a broad and purposive approach to interpretation, as urged by the IEBC, the only conclusion was that only ‘valid’ (i.e., unspoiled) votes should be included in the runoff calculation.Footnote 45 It is unclear why a broad and purposive approach to interpretation supports the view that ‘all votes’ does not include spoiled ballots. The Court only briefly cited a judgment of the Seychelles Court of Appeal in support, without evaluating or even rehearsing its reasoning.Footnote 46
Read as a whole, the effect of the judgment is to ‘insulate both the IEBC and the candidate which it declares to have won in the first round from effective challenge in almost all cases.’Footnote 47 Although the Court offered a handful of additional reasons not to intervene in the election results, it is difficult not to see them as anything other than post-hoc justifications for an a priori determination that the Court should not intervene.
Judicial confidence: the 2017 election
While in 2013 the Court was satisfied with the IEBC's evidence and arguments, in 2017 it took a much more inquisitorial approach. Early on in the proceedings the Court ordered an independent audit of the presidential election forms – unimaginatively named Form 34A, Form 34B and Form 34C – and the electronic process by which they were generated. Section 39(1C) of the Elections Act requires election officers at each polling station to fill in the results on Form 34A and transmit a scanned image of the form to the constituency tallying centre (CTC). The returning officer at the CTC is then to verify and tabulate the results from the various polling stations in the constituency and generate form 34B. The results are then to be sent to the national tallying centre (NTC) where the chairperson of the IEBC must verify results and generate Form 34C.
Relying on the audit report, the Supreme Court confirmed a handful of irregularities. We highlight just two here. First, the IEBC announced the results of the election before its own electronic system had confirmed all the votes. Election results had been transmitted from polling stations to tallying centres without scanned images of Form 34A,Footnote 48 and results were thus improperly tabulated on Form 34B. In turn, the results on the IEBC website differed from those on the various Forms 34B and 34A.Footnote 49 The IEBC admitted in court that it had declared the election results before receiving authentic Forms 34A from polling stations in respect of more than 3.5 million ballots.Footnote 50
Second, the Court accepted evidence casting doubt on the authenticity of the forms themselves. Forms 34A and 34B have security features consisting of serial numbers, barcodes, and the IEBC's watermark.Footnote 51 However, as the audit discovered, 59 out of 290 Forms 34B were missing some or all of these security features. Five of them were not signed. Many of the Forms 34A were not stamped or signed, and many of them were questionable photocopies.Footnote 52 The missing signatures were a great puzzle to the Supreme Court, which noted that a signature is a simple and easy way for an electoral officer to attest to the accuracy of results.Footnote 53
Taken together, the Court held that these problems raised reasonable doubts as to whether ‘the election can be said to have been a free expression of the will of the people’.Footnote 54 In taking the unprecedented step of nullifying the presidential election, the Court concluded that irregularities in the electoral process or non-compliance with constitutional and statutory requirements need not affect the outcome of the election in order for them to render an election unfair. Rather, irregularities and non-compliance are unfair in and of themselves. ‘Where an election is not conducted in accordance with the Constitution and the written law,’ the Court held, ‘then that election must be invalidated notwithstanding the fact that the result may not be affected’.Footnote 55
In reaching this conclusion, the Court rejected the IEBC's arguments that unfairness flows only from irregularities significant enough to alter the outcome.Footnote 56 We take no position on this particular point of electoral law, but we want to emphasise that the Court engaged in the debate in the first place. It did not simply defer to the IEBC's submissions without explaining its reasons for doing so, as it did in 2013.Footnote 57 In 2017, the Court concluded that nullifying the election was almost inevitable under the weight of evidence:
[I]t is our finding that the illegalities and irregularities committed by the [IEBC] were of such a substantial nature that no Court properly applying its mind to the evidence and law as well as the administrative arrangements put in place by the IEBC can, in good conscience, declare that they do not matter.Footnote 58
The Court's response in the 2017 election is consistent with our view that when it comes to functions constitutive of voting rights – like tabulating and counting votes – courts should adopt an appellate posture and reason through the facts and the law themselves. Courts should come to their own conclusion as to whether an election was fair, rather than asking only whether an EMB's view as to the election's fairness meets an administrative law standard of reasonableness.
If the Kenyan Supreme Court had engaged in traditional administrative law review in 2017, it might have been compelled to uphold the election. Without a de novo hearing on the merits and searching scrutiny of the evidence relating to the transmission and authenticity of voting forms, a conclusion that the 2017 elections had been fair does appear to be reasonable. Indeed, international observers from the EU, the AU, the Commonwealth and the East African Community had overwhelmingly found the elections to be generally free.Footnote 59
More to the point, the Court's assessment was the first time that the IEBC's decision was scrutinised at all, by anyone other than the IEBC itself. Without judicial scrutiny on the merits, it is not clear to whom voters concerned about the integrity of an election could turn. The constitutive nature of the IEBC's function here, in breathing life into democratic rights, justifies judicial intervention when it comes to reviewing complaints about election results.
The distance between 2013 and 2017
On our model, the Supreme Court should have looked as deeply into the substance of the electoral petitions in 2013 as it did in 2017. Instead, the 2013 Court accepted the IEBC's arguments about deference and satisfied itself with the IEBC's assurances that the election was fair. Although our model is prescriptive rather than descriptive, there is one sense in which it does explain the decisions: the Court simply took the wrong approach in 2013.
In this section of the paper we consider why the Court's decisions are so far apart. To do so, we build on the Court's own admission in the 2013 judgment that it did not have enough ‘political capital’ to spend.Footnote 60 The 2017 election, evidently, was one of the ‘rare occasions’ when it was appropriate for the Court to dip into its savings. As we have argued throughout this paper, however, the dilemma presented by the 2013 and the 2017 election petitions was essentially the same. It is not so much that the ‘occasion’ was different in 2017, but rather that the Supreme Court had more political capital to spend.
The Court's increased political standing, and its confidence in that standing, is in large part due to the process of vetting judges and magistrates. Before 2010, the Kenyan judiciary was widely perceived to be aligned with, if not entirely captured, by the ruling party. Raila Odinga did not even bother to approach the courts in 2007 after the incumbent president, Mwai Kibaki, was declared to have won the election. Part of the transition to the new constitutional order was therefore what has been referred to as the ‘radical surgery’ of the Kenyan judiciary.
Acknowledging that the judiciary had a deep-seated problem with systemic corruption, the drafters of the Constitution agreed that every judge and magistrate who held their position before the effective date of the 2010 Constitution would have to reapply for their job and undergo a vetting process governed by statute.Footnote 61 If the Vetting of Judges and Magistrates Board found that they had failed to act with integrity, fairness and honesty prior to 2010, they could be removed from the position. The process saw the dismissal of eleven judges and fourteen magistrates.Footnote 62 Reasons given for their dismissal included lack of impartiality, incompetence, negligence, conduct unbecoming of a judicial officer and being aware of human rights violations but doing nothing about them.Footnote 63
Separately, with the growing recognition in the wake of post-election violence in 2007 that effective electoral dispute resolution could play a key role in national stability, Chief Justice Willy Mutunga formed the Judiciary Working Committee of Election Preparations (JWCEP) in 2013. The JWCEP (now the Judiciary Committee on Elections) was initially comprised of nine judges and magistrates from the Supreme Court, Court of Appeal, High Court and Magistrates Courts.Footnote 64 Its mission was to strengthen and improve the Judiciary's capacity to resolve election petitions in a ‘people-focused’ and expeditious manner. To do so, it aimed to design and implement a programme to build the capacity of judges with respect to the adjudication of electoral disputes.Footnote 65 While the vetting process worked to increase public confidence in the judiciary as a whole, the JWCEP's work helped to restore public confidence in the judiciary's role as an arbiter in electoral disputes.
During the same period from 2013 to 2017, the IEBC's public standing dropped. First was the ‘Chickengate’ scandal. Between 2015 and 2017, an investigation by the Ethics and Anti-Corruption Commission (another constitutional fourth-branch institution) revealed details of an arrangement in which a British printing firm paid bribes to IEBC officials in return for ballot-printing contracts (the bribes were referred to as chickens in their communications).Footnote 66 Several IEBC officials were arrested and others resigned.Footnote 67 By August 2016, a year before the next general election was scheduled to be held, there was essentially no electoral commission.
Second, the selection of new commissioners was plagued by accusations of favouritism. The President's nomination of Wafula Chebukati as chairperson was particularly controversial, as he had lower qualifications compared to other candidates.Footnote 68 The new Commission took office in January 2017, with Chebukati as chairperson, with less than seven months to go until the election. That timeline would be a challenge even for a qualified and competent commission, but with relatively inexperienced and apparently partisan commissioners and a recent history of corruption, the IEBC commanded little public trust leading up to the 2017 election.Footnote 69
With the IEBC slipping in public trust and the courts newly ascendant, the political cost of judicial intervention in electoral outcomes was greatly reduced. In terms of political capital, the 2017 Court was much better positioned than the 2013 Court to subject the IEBC to appellate scrutiny.
Judicial responses to regulative functions
While we think that the Supreme Court made a mistake in 2013 by refusing to engage in substantive review of the IEBC's constitutive functions, the courts have at times been appropriately deferential to the IEBC's regulative functions, and in particular to its dispute resolution functions under Article 88(4)(e) of the Constitution.
It bears noting here that this provision specifically excludes disputes relating to the declaration of results from the IEBC's dispute resolution jurisdiction. Any complaints about the outcome of an election will have to be presented to the courts in the first instance. This is consistent with our functional argument in two ways. First, the declaration of electoral results is more constitutive of voting rights than regulative of them. Determining who will be your next president (or prime minister or legislative representative) is in many ways the purpose of a person's right to vote. Counting votes and in turn declaring a winner is what makes a person's vote meaningful: these functions are constitutive of the right to vote, and are the kind of function that courts should be involved in overseeing. Second, without judicial oversight of the IEBC's determination of electoral results, there would be no forum for independent secondary oversight of elections.
The dispute resolution authority that the 2010 Constitution confers on the IEBC is thus limited to regulative matters. We consider here a handful of court judgments illustrating how the courts have approached review of these regulative matters in a way that is consistent with our model's normative prescriptions. In the 2013 High Court case of Diana Kethi Kilonzo & Another v Independent Electoral & Boundaries Commission & 10 Others,Footnote 70 for example, the petitioners challenged the IEBC Dispute Resolution Committee's holding that the first petitioner was ineligible to contest a senatorial by-election for Kenya's Makueni County. The DRC made this initial determination because it could not trace the candidate's registration in the biometric voter registration database, and the serial number of the voter registration slip she submitted turned out to have been used before – by former President Mwai Kibaki, no less.Footnote 71
The Court affirmed that under Article 165(3) of the Constitution, it had supervisory jurisdiction over the IEBC and, by extension, the DRC, but only with regards to ensuring that ‘the Constitution is obeyed and respected by tribunals and other bodies’.Footnote 72 That is, the Court saw its role in this case as confined to policing the boundaries of the IEBC's lawful powers. The Court therefore adopted a position of deference to the IEBC:
The Constitution allocates certain powers and functions to various bodies and tribunals. It is important that these bodies and tribunals should be given leeway to discharge the mandate bestowed upon them by the Constitution so long as they comply with the Constitution and national legislation.Footnote 73
The Court then held that the DRC had the authority to hear and determine the complaint in accordance with the powers granted to the IEBC by Article 88(4)(e) of the Constitution.Footnote 74 In dismissing the petition, the Court observed that there was no indication that the first petitioner had been denied a fair hearing.Footnote 75
A 2016 case has a similar thrust. In Moses Mwicigi v Independent Electoral and Boundaries Commission,Footnote 76 the Supreme Court affirmed a decision of the DRC that a complaint was improperly put before it (i.e., that the DRC did not have jurisdiction to hear the complaint). The High Court had agreed, but the Court of Appeal set aside the DRC's decision and substituted it with its own order. The Supreme Court upheld the High Court's decision and restored the DRC's original order, castigating the Court of Appeal in a passage that speaks directly to the importance of allowing a constitutionally mandated fourth-branch institution to fulfil its regulative functions:
To allow an electoral dispute to be transmuted into a petition for the vindication of fundamental rights under Article 165(3) of the Constitution, or through judicial review proceedings, in our respectful opinion, carries the risk of opening up a parallel electoral dispute-resolution regime. Such an event would serve not only to complicate, but ultimately, to defeat the sui generis character of electoral dispute-resolution mechanisms, notwithstanding the vital role of electoral dispute-settlement in the progressive governance set-up of the current Constitution.Footnote 77
We mention a final case because it involves regulative functions performed by a non-constitutional institution. Kenya's Political Parties Dispute Tribunal (PPDT) is established by the Political Parties Act, and its mandate is to resolve disputes within political parties, between political parties and between independent candidates and political parties.Footnote 78 The PPDT's functions are regulative because they safeguard democratic rights threatened by other actors – in this case, political parties.
In Bob Micheni Njagi v Orange Democratic Party,Footnote 79 a complainant approached the PPDT alleging irregularities in the Orange Democratic Party's nomination process. In reviewing the PPDT's decision, the High Court considered only whether the PPDT had properly considered the relevant law and the information before it and arrived at a reasonable conclusion. The court deferred to the PPDT's decision because it had been reasonable, rather than deciding for itself what the answer to the dispute might have been. In regulative cases like this, courts need do no more.
Conclusion
The distinction we draw between IPDs’ constitutive and regulative functions suggests when it might be justifiable for courts to evaluate the substantive merits of an IPD's decision, and when they should confine themselves to more traditional administrative law review. The model is compatible with Pal's equivalency thesis because it ensures a similar degree of scrutiny for each of the four constitutional branches of government. IPD decisions will be subject to substantive review when their functions are constitutive of rights, and when it is thus the IPD itself whose conduct impacts constitutive rights. The executive and legislative branches are subject to similar kinds of judicial scrutiny.
When government (or sometimes private) conduct affects a person's ability to exercise rights they already hold and an IPD is called on to protect or vindicate those rights, the IPD's role is regulative. An IPD's regulative oversight role in these situations should be subjected to administrative law review that looks only for reasonableness, lawfulness and procedural fairness. In the case of EMBs like Kenya's IEBC, its dispute resolution mechanisms allow it to perform regulative functions that safeguard the exercise of democratic rights throughout the electoral process.
Our model is also consistent with Tushnet's assessment that high courts are unlikely to give up substantive oversight over highly consequential fourth branch decisions. Seeing IPD decisions in terms of their constitutive and regulative impact suggests why it is justifiable for courts to retain substantive oversight when IPDs perform constitutive functions.
Kenya's experience makes it clear, however, that whatever normative prescriptions our model makes, courts of review will likely be influenced by the broader political landscape. How our model suggests a court should act, or would be justified in acting, is not always how the relative institutional strengths of courts and IPDs will allow it to act. Constitutional principle and political pragmatism do not always align.
Author Biography
Richard Stacey, Associate Professor, Faculty of Law, University of Toronto.
Victoria Miyandazi, Advocate of the High Court of Kenya. Lecturer, School of Law, University of Embu.
