On 13 May 2021, the High Court of Kenya handed down a landmark judgment striking down the “Building Bridges Initiative”, a set of proposed amendments to the Kenyan Constitution, on grounds – inter alia – of violating the Constitution’s basic structure [“the BBI Judgment”]. At the time, I had analysed the judgment in this blog post. The High Court’s judgment was appealed to the Court of Appeal, and argued before a seven-judge bench, over four days, between June 29 and July 2. Judgment has been scheduled for August 20.
The Kenyan constitutional courts are unique in that oral arguments are live-streamed, broadcast live by major television networks, and available on YouTube (with an open comments section!). I was thus able to watch the arguments, which raised a range of fascinating issues, from the distinction between direct and indirect democracy (drawn from the Kenyan Constitution’s very particular amendment provisions), the identity of “the People”, the distinction between top-down and bottom-up constitutionalism, the meaning and purpose of public participation, and so on. Perhaps what made the arguments even more interesting was that, having been enacted in 2010, the Kenyan Constitution is a relatively young document, and not yet bowed down by layers of encrusted precedential sediment. For this reason, counsel on both sides were able to make arguments on first principle, presenting the overlaps between concrete constitutional practice and the philosophy of democratic constitutionalism, in a distilled form that we don’t often get to see.
Here, I will examine some of the major arguments made over the course of the four-day appeal hearing. The first two days of the hearing were occupied by the appellants who were challenging the High Court’s decisions. These included the Attorney-General’s Office, the BBI Secretariat, the Independent Electoral and Boundaries Commission (whose powers – and their alteration – were central to the High Court’s judgment), the Hon. Raila Odinga (the opposition leader), and HE Uhuru Kenyatta, President of Kenya. I will not here examine all the arguments raised; that would be an impossible task, apart from the fact that some of the issues are beyond the scope of a comparativist’s ability to engage with. I will, however, discuss three issues that, in my view, are at the heart of the appeal.
Perhaps what made the arguments even more interesting was that, having been enacted in 2010, the Kenyan Constitution is a relatively young document, and not yet bowed down by layers of encrusted precedential sediment.
My usual caveat: I am an outsider, and I approach this post with an awareness of my limited knowledge. While I will write as honestly as I can, there is undoubtedly much context that I do not know, and much that I may get wrong; for those trespasses, an apology, in advance.
Who Are The People?
First, Appellants challenged the High Court’s finding that the Kenyan Constitution had a basic structure, which could only be amended through the exercise of a primary Constituent power that – in effect – lay outside the Constitution. Recall that in its judgment, the High Court had not held that the Kenyan Constitution’s basic structure was beyond the power of amendment altogether (thus making the Kenyan basic structure doctrine entirely different from its Indian variant, and indeed, unique in the annals of comparative constitutionalism). Rather – on an analysis of Kenyan constitutional history – the High Court held that in order to amend the basic structure, the conditions around the founding of the 2010 Constitution had to be recreated (primary constituent power). This meant, in concrete terms, a four-step procedure involving civic education, public participation, the convening of a Constituent Assembly, and a referendum.
It was therefore a little surprising to note the amount of time that Appellants spent distinguishing Kesavananda Bharati v State of Kerala, or arguing for its inapplicability to Kenya. As I have mentioned above, the basic structure doctrine of Kesavananda, and the basic structure doctrine of the BBI judgment, are chalk and cheese. In particular, Appellants spent considerable time arguing that Kesavananda was a response to Parliamentary abuse of amending power, while the Kenyan Constitution explicitly envisages a role for the people (via the Popular Initiative method under Article 257) in the amending process. This is correct. But it is also, with respect, irrelevant. The distinction would have made sense if the Kenyan High Court had placed the basic structure out of bounds altogether; however, as we have seen, it did not – arguing only for deeper public participation than is provided in the Kenyan Constitution via the popular initiative method. That argument, however, needs to be tackled on its own terms, and not with reference to Kesavananda and parliamentary supremacy.
That brings us to how Appellants did attempt to tackle it on its own terms. Now here is where things get interesting. Let us go back for a moment to the amendment provisions in Kenya’s Constitution (Articles 255 – 257). Recall that Article 256 prescribes the familiar parliamentary route for certain constitutional amendments. Article 257, however, provides an alternative route known as the “popular initiative”, which requires (a) signatures of one million registered voters to initiate; (b) approval by a majority of county assemblies; (c) passage by simple majority in both Houses of Parliament, and (d) if pertaining to one of ten critical categories spelt out in Article 255, a referendum. Interestingly, therefore, the method under Article 257 is a combination of direct and representative democracy: it has elements of direct democracy in its initiation and completion (in cases of a referendum), and also elements of representative democracy through the process (ratification by county assemblies and Parliament).
It was therefore a little surprising to note the amount of time that Appellants spent distinguishing Kesavananda Bharati v State of Kerala, or arguing for its inapplicability to Kenya
In this context, Appellants argued that the participation of the people is already provided for under the Constitution, when it comes to the question of amendments. Consequently, there was no need for a separate basic structure doctrine that was itself based on public participation, as that issue had already been resolved within the Constitution. To supplement this argument, counsel referred to Article 1(1) of the Kenyan Constitution, which states that “all sovereign power belongs to the people of Kenya and shall be exercised only in accordance with this Constitution.” Stressing upon the phrase “this Constitution”, Appellants argued, therefore, that popular sovereignty did not – and could not – exist outside the Constitution. It was the Constitution that – through Article 257 – facilitated the exercise of popular sovereignty through the popular initiative amendment process.
But underlying this textual point is a deeper normative claim, which goes to the heart of constitutionalism: who are the People? Effectively, Appellants argued that “the People” had no existence – or identity – outside of the Constitution; quite literally, the People are “constituted” by the Constitution. Consequently, in holding that the basic structure could only be amended by an – effectively – extra-Constitutional process initiated – and led – by the People, the High Court had misconstrued not only the text of the Constitution, but the basic premises of constitutionalism itself.
At first blush, the argument sounds powerful. I believe, however, that in the final analysis, it fails to persuade. I would submit, with respect, that in focusing on Article 1(1), Appellants inadvertently hoisted themselves on their own petards. This is because the phrase “this Constitution” begs the anterior question, which is at the heart of the basic structure doctrine: at what point does “this Constitution” cease to be “this Constitution”, and becomes something else? It is here that Kesavananda Bharati actually becomes relevant – because what many people (especially in its home country) tend to forget is that before it was anything else, Kesavananda Bharati was a textual judgment. The majority judges took great pains to parse the meaning of the word “amend”, and took equal pains to explain how it differed from “repeal” or “abrogate” or “destroy”. To put the point at its simplest: this Constitution – whose entire structure is based on the principle of (say) representative democracy, is no longer this Constitution if all those provisions are replaced with a clause turning the State into a monarchy. Indeed, the bench repeatedly put these hypotheticals to the Appellants’ counsel – without (in my view) a satisfactory answer.
Indeed, Respondents’ counsel made this point on multiple occasions. The clearest exposition of it is to be found from around 55 minutes to 1 hour during the first session of Day 3: counsel distinguished between the concepts of “amendment” and “repeal”, and then further distinguished between two kinds of repeal: express repeal (such as when Kenya’s Independence Constitution was “retired”, in advance of the 2010 Constitution coming into force); but also, repeal by irreconcilable differences, where although a Constitution is not formally repealed, it is nonetheless repealed in effect, by introducing into it principles or provisions that are fundamentally irreconcilable with its core identity.
To this, I would add two brief points. The first is that taking the argument to its logical conclusion, Article 1(1) of the Kenyan Constitution, far from supporting Appellants’ case, is fatal to it. Because if it is true that (a) Article 1(1) only authorises the exercise of popular sovereignty within this Constitution, and (b) that this Constitution is no longer this Constitution if its basic structure is altered, it must necessarily follow that (c) if the People wish to alter the basic structure, they must act outside the Constitution. Indeed, it cannot be otherwise. And if this is conceded, then the High Court’s judgment is not only arguably but correct, but indeed, the only correct outcome. Because not only did the High Court correctly rule that alteration of the basic structure must take place outside of the Constitution, it also set out the logical corollary: that the manner of doing so must resemble, as closely as possible, the manner of the writing of the Constitution. This is so because an alteration of the basic structure amounts, in effect, to re-writing the founding document, and for that, such an activity needs to be conducted within the procedural confines of the manner in which its writing took place.
Counsel on both sides were able to make arguments on first principle, presenting the overlaps between concrete constitutional practice and the philosophy of democratic constitutionalism, in a distilled form that we don’t often get to see.
The second point is that on multiple occasions, Appellants’ counsel expressly conceded that the Constitution does have a basic structure. Counsel attempted to make light of the point by arguing that every form of government – including theocracies and one-party totalitarian States – have basic structures. Counsel then argued that the existence of a basic structure doesn’t necessarily imply the existence of the basic structure doctrine. However, it is in precisely that disjunction where the problem lies: if you accept that the Constitution has a basic structure – i.e., a core set of principles that constitute its identity – then the burden falls upon you to show how an alteration of the basic structure, and thus an alteration of constitutional identity nevertheless implies that “this Constitution” at the beginning of the process is still “this Constitution” at the end of it. I believe that this was a burden Appellants did not discharge.
The Njoya Debate
I would suggest that the above argument is also relevant to answering a question that was hotly debated on both sides of the bar during the hearings: that is, the interpretation of the Kenyan High Court’s 2004 judgment in Njoya v Attorney-General. In Njoya – which involved a challenge to the then-ongoing constitutional review process – the High Court had accepted Kesavananda Bharati, and held that under the (now-retired) Constitution, the Kenyan Parliament had no power to abrogate or repeal the Constitution. Now naturally, the Respondents strongly argued that Njoya went in their favour. However, Appellants argued equally strongly that read closely, Njoya actually supported their case. The reason for this was that in Njoya, Ringera J categorically stated that (a) Parliament had no power to abrogate the Constitution, (b) that this power belonged to the People of Kenya, acting in their sovereign capacity, and (c) that the People of Kenya were entitled to a referendum on any new Constitution (paragraphs 32 & 33, Njoya). Now if you look at Articles 255 read with 257, an amendment by popular initiative – that seeks to alter one of the ten categories set out in Article 255 (such as the territory of Kenya or the functions of Parliament) must be put to a referendum. Appellants therefore argued that the dictum of Njoya had already been incorporated into the 2010 Constitution. Indeed, the bench’s question to Mr. Nelson Havi, Respondents’ lead counsel, at around 48 mins of the first session of Day 3, went to the heart to the issue: did not the 2010 Constitution enact the very requirements outlined in Njoya?
I believe that the answer is in Njoya itself, and it is the same answer as that set out above. In paragraph 29, Ringera J wrote that popular sovereignty is “the basis of the creation of the Constitution and it cannot therefore be conferred or granted by the Constitution” – although, of course, it could be juridically recognised by it. These lines, I submit, answer the bench’s question about the referendum provisions under Article 257: regardless of what Article 257 says about a referendum, ex hypothesi, the power to create a Constitution cannot lie within it. Now when you combine this with Njoya’s holding in paragraph 61 accepting the correctness of Kesavananda and holding that “amendment” does not mean “abrogation” or “repeal”, you reach the inescapable conclusion that a Constitution’s amendment provisions – whether or not they contemplate referenda – do not extend to the repeal of this or the creation of a new Constitution, something that can only be done through an external exercise of popular sovereignty. And that was what the High Court recognised.
The Discontents of the Presidential System
The second major issue was with respect to the identity of the initiator of the BBI Amendment(s). Articles 257(3) and (4) of the Kenyan Constitution contemplate the existence of the “promoters” of an amendment by popular initiative. Formally, the promoters of the BBI Amendment(s) were two parliamentarians named Dennis Waveru and Junet Mahomed. However, it was strongly urged before – and accepted by – the High Court that the actual moving force behind the Amendment(s) was the President of Kenya, Uhuru Kenyatta. The factual dispute continued before the Court of Appeal, and I do not intent to address it here.
However, it was also argued by Appellants that even if it was the President who was behind the BBI Amendment(s), there was nothing wrong with it; in other words, just like any other registered voter, high State officials could also initiate amendments by popular initiative. In particular, Appellants argued that (a) there was no express bar under Article 257 that forbade the President from engaging in amendment(s) by popular initiative; (b) that explicitly, the President continued to retain political rights under the Constitution even on formal investiture, including, for example, the right to vote; (c) that as a matter of fact, even if the President had been acting in this case, he had been acting in his private capacity; and (d) that the entire purpose of Article 257 was to prevent legislative bottlenecks in a Presidential system. It was meant to deal with a situation where the political party with a majority in Parliament was opposed to the elected President’s agenda, and was stymying him. In such a situation of gridlock, Article 257 gave the President an avenue to to bypass Parliament and go directly to the People.
In this context, Appellants argued that the participation of the people is already provided for under the Constitution, when it comes to the question of amendments. Consequently, there was no need for a separate basic structure doctrine that was itself based on public participation, as that issue had already been resolved within the Constitution.
At its deepest level, then, the argument was one about constitutional design. Appellants argued that the Kenyan Constitution did two things: it opted for the Presidential system, with its known problem of veto points and bottlenecks – and then also solved that problem through Article 257, by essentially saying “President + People > Parliament“.
Now there are a few points here. The first is that – as the bench acutely pointed out – Article 255(3)(b) of the Constitution states that the amendment procedure under Article 257 is by “the people and Parliament”. It therefore follows that the Constitution itself draws a distinction between “the People”, and constitutional bodies (such as the Parliament), even though – of course – those constitutional bodies are helmed by members of “the People”. It follows ipso facto, therefore, that much like “the people” and “Parliament” are distinct entities for the purposes of a constitutional amendment, “the people” and “the Presidency” must likewise be so.
Is this dispositive? No. However, what it does show is that textually, Articles 255 – 257 sustain the High Court’s interpretation. Whether it is the correct interpretation – given that the text is ambiguous – is of course a normative question, internal to the Kenyan Constitution. And it is here that the principles of representative and direct democracy come into tension with each other. As I have pointed out above, Article 256 provides the familiar amendment path via representative organs (Parliament), while 257 consciously combines elements of representative and direct democracy, and – specifically – envisages direct democracy at both the beginning and the end of the process.
I would submit, with respect, that this carefully balanced scheme of Article 257 would be nullified were representative organs allowed to initiate the process. Such an interpretation of Article 257 would undermine the fact the bottom-up character of direct democracy that the provision seeks to preserve. During the opening session of Day 3, at around 1 hour 6 minutes, Respondents’ counsel put the point perfectly when they argued that by definition, when the people delegate authority (as in representative democracy), the delegate no longer counts as part of ‘the people’ for the purposes of the exercise of popular sovereignty (as in direct democracy, envisioned by Article 257).
The bench then asked – correctly – whether it would not be possible for the President to “wear two hats” – i.e., remain the President, while initiating a popular initiative process as a private citizen. The answer – which was provided a little later in the day – was that ‘yes, she can, but she needs to formally give up the trappings of office while doing so.’ This, to me, makes eminent sense: the entire point of Article 257 – and the carefully wrought distinction between direct and representative democracy – would be defeated if the President could say “I am President, but I ask for your signature on this popular initiative as a private citizen.” With respect, that distinction could be sustained only by denying reality.
As I have mentioned above, Appellants repeatedly argued that there was no express bar on the President’s powers under Article 257. That is true; but again, going back to Kesavananda – and much more recently, to the UK Supreme Court’s prorogation judgment in Miller v The Primer Minister, limitations upon powers exercised by constitutional authorities are not only express, but – in many cases – implied. In Miller, it was held that when you have two constitutional principles, the exercise of power under one is (impliedly) limited at the point at which the said exercise would frustrate or impede the operation of the other. That is exactly what is happening here: the President’s powers under the Constitution are limited by the operation of other constitutional principles; in the case of Article 257, that principle is the balance between direct and representative democracy in the popular initiative process.
The Nature of Referenda
The third issue also concerned the interpretation of Article 257. The BBI Amendment(s) had been pushed as a package deal, containing as many as seventy-four proposed amendments to the Kenyan Constitution. The High Court had nixed this in no uncertain terms, holding that in an amendment by way of popular initiative, every amendment had to be presented distinctly, and put through the process required by Article 257, separately. Appellants’ response to this holding was primarily textual, noting that Article 257 used the word “amendment”, and not “amendments.” Thus, an “amendment” could take the form of a proposed bill – that amended multiple provisions of the Constitution – but for the purposes of Article 257, the was to be treated as an amendment, in the singular.
Now it should be clear that once again, the text does not determine the issue. It could easily well be argued – and indeed, was argued – that the very use of the singular “amendment” actually means that you can only present one amendment at a time via the popular initiative. For interpretive support, Respondents’ counsel also pointed to Kenya’s Elections Act, Section 51(2) of which provides that “where there is more than one referendum question, persons intending to campaign for or against each referendum question shall, on application to the Commission, form one national referendum committee each and one committee each in every constituency for each referendum question.”
At first blush, the argument sounds powerful. I believe, however, that in the final analysis, it fails to persuade. I would submit, with respect, that in focusing on Article 1(1), Appellants inadvertently hoisted themselves on their own petards. This is because the phrase “this Constitution” begs the anterior question, which is at the heart of the basic structure doctrine: at what point does “this Constitution” cease to be “this Constitution”, and becomes something else
Textual ambiguity, once again, brings us into the domain of the normative – and the normative issue here goes once again to the purpose of referenda, and the provision of direct democracy in a Constitution. The core argument against referenda – as we well know – is that they boil down extremely complex and nuanced questions into a simple yes/no binary. In effect, therefore, they distort the expression of public opinion by simplifying an issue to a point where a yes/no vote doesn’t accurately reflect the range of views that people hold. Now, I would submit, with respect, that given that referenda already have this problem with respect to being accurate vessels for the expression of direct democracy, a package deal referendum exacerbates that precise problem even more (the reasons are easy to see). It therefore follows that if the Court is faced with two equally legitimate textual interpretations of the same provision, one of which advances the cause of direct democracy, and the other potentially impedes it, given that the purpose of the section is to facilitate direct democracy, it should choose the former interpretation. In this case, that would be to uphold the High Court’s ruling that a popular initiative can only be for one amendment at a time.
The above analysis has only scratched the surface of the immensely complex and nuanced arguments that I heard. Some of the fascinating arguments I have not dealt with here include, for example, the question of which side bears the burden in a dispute over whether the popular initiative involved adequate public participation or not (I would suggest that if public participation is a fundamental premise of the Kenyan Constitution, the presumption must be that official acts do not acquire a constitutional seal of approval until it is affirmatively demonstrated that public participation has taken place; think of it as a precondition); the role of constituencies in distributing resources, power, and political patronage, and why, therefore, boundary alteration affects the basic structure; whether the popular initiative starts from the moment signatures are collected, or does the public participation requirement kick in before; and so on. Ultimately, what I found most striking is that for a lot of these questions, the Constitution does not provide a textual answer. For a lot of these questions, both sides presented an interpretation that, textually, was plausible; ultimately, therefore, this ruling will turn upon what principles the Court identifies as underlying these provisions, and which interpretation it believes better furthers those principles.
A few parting remarks on the oral arguments themselves. I loved watching them: there was wit, humour, seriousness, and a whole range of styles of advocacy, with consistently high quality throughout. The timekeeping was impressively strict: a big timer showed counsel their countdown, the bench cut them short when their time was up, and barring a few occasions where I thought that counsel genuinely needed more time to develop their argument, I felt that the quality of advocacy was raised by the fact that counsel had to put their point across in limited time. The questions from the bench were sharp, focused, and at times, almost frightening in how accurately they identified weak spots in the argument. And then, away from all this, I was struck by the freedom with which people were commenting and arguing with each other in the YouTube comments section, and openly poking fun at the lawyers, the President, and even the court. As far as I could tell, none of this had any perceivable impact on “justice”. Perhaps judges in other jurisdictions may take a leaf out of this YouTube book, and treat irreverence and critique with a smile.
Finally, as an outsider, I was struck by the civility with which the arguments were conducted. By all accounts, this is amongst the most – if not the most – important constitutional case under Kenya’s new Constitution, and the High Court’s judgment was a massive setback for both President Uhuru Kenyatta, and the Hon. Raila Odinga. However – barring a few good-natured jabs back and forth, I did not see personal invective. Imagine watching the State’s lawyers argue over two days, and not once insinuate that the other side is a hotbed of anti-nationals, publicity-seekers, “lobbyists”, and a threat to sovereignty and national security?