Who owns a “popular initiative” to amend a constitution? Attorney General & 2 others v Ndii & 79 others, the Supreme Court of Kenya’s 2022 Building Bridges Initiative judgment, turned that phrase into a practical boundary on public power. The case did not hold that Kenya’s Constitution cannot be amended in fundamental ways. Instead, it asked whether constitutional change had travelled through the route the Constitution actually provides, whether a route reserved for popular initiative could be initiated through presidential authority and state machinery, and whether materially new proposals had received the public participation that constitutional change demands.

The BBI process and why its origin mattered

The dispute grew out of the political process known as the Building Bridges Initiative. After the “handshake” between President Uhuru Kenyatta and Raila Odinga, the President appointed an advisory taskforce and later a steering committee. Their work produced the Constitution of Kenya (Amendment) Bill, 2020, a wide-ranging package of proposed amendments. The proposal was advanced as a popular initiative under Article 257, signatures were collected and submitted for verification, and the process moved through county assemblies toward a possible referendum.

Several petitioners challenged the process. The High Court invalidated important parts of it in May 2021, and the Court of Appeal largely upheld that result in August 2021. By the time the dispute reached the Supreme Court, the case included questions about the so-called basic structure doctrine, the President’s role in a popular initiative, public participation, the proposed allocation of seventy new constituencies, the composition of the Independent Electoral and Boundaries Commission, referendum questions, and presidential immunity. Because all seven Supreme Court judges wrote separately, reading the final decision requires attention to the majority on each issue rather than treating every powerful passage as a holding of the Court.

Article 257 is a route with an owner and a sequence

Article 257 allows a constitutional amendment to be proposed by a popular initiative signed by at least one million registered voters. The initiative may begin as a general suggestion or a formulated draft Bill. Its promoters deliver the draft and supporting signatures to the IEBC for verification. If the constitutional requirements are met, the Bill proceeds to county assemblies. Approval by a majority of county assemblies sends it to Parliament, and Article 257 then determines whether parliamentary passage is sufficient or whether the proposal must go to the people in a referendum, including where it concerns matters protected by Article 255(1).

The sequence matters because it allocates constitutional responsibility. The promoters promote; the IEBC verifies; county assemblies consider; Parliament performs its assigned role; and the people decide by referendum when the Constitution requires it. A public body cannot make the procedure legitimate merely by giving a constitutionally recognised label to a process that began or developed in a materially different institutional form.

The President cannot initiate the citizens’ route

A majority of the Supreme Court held that the President cannot initiate a constitutional amendment through Article 257’s popular-initiative route. The constitutional presidency carries public authority, staff, institutional influence and access to state resources. It cannot be treated as an ordinary private citizen for purposes of originating a route designed for direct popular sovereignty. The majority also found that President Kenyatta had in fact initiated the BBI amendment process. On that basis, the proposed Bill could not lawfully proceed as an Article 257 popular initiative.

This holding is best understood as a rule about constitutional capacity and route, not a gag on presidential political speech. A President may participate in public debate about constitutional reform and public institutions may perform functions the Constitution assigns to them. The problem is using the institutional presidency to originate a process and then characterising it as the people’s Article 257 initiative. The route protects a space in which constitutional change is supposed to be initiated from outside the state machinery that ordinarily exercises delegated sovereign authority.

The Court rejected the lower courts’ basic-structure formula

The BBI litigation became internationally discussed because the High Court and Court of Appeal accepted versions of the basic structure doctrine and a four-stage process involving civic education, public participation, a constituent assembly and a referendum for certain fundamental changes. The Supreme Court did not adopt that formula. By a six-to-one majority on the formal question, it rejected the basic structure doctrine as formulated below and held that the four sequential stages were not an additional constitutional amendment pathway imposed by Kenyan law.

That distinction prevents a common misstatement of Ndii. The Supreme Court did not say that constitutional structure is legally irrelevant. It said that Chapter Sixteen already creates a deliberate system of amendment through Articles 255, 256 and 257 and that courts should not add a free-standing fourth procedure merely because they consider some provisions especially fundamental. Arguments about constitutional identity must therefore be tied carefully to the text, amendment route, entrenched referendum requirements, institutional allocations of power and other binding constitutional limits.

The seventy constituencies show why participation must match the final proposal

The proposed Second Schedule allocated seventy new constituencies. The Supreme Court unanimously held that it was unconstitutional. The schedule had been added late and had not itself been subjected to public participation, and its allocation of constituencies also intruded on the constitutional boundary-delimitation role assigned to the IEBC under Article 89. This gives the case a practical lesson that extends beyond BBI: consultation on a general reform idea does not automatically validate a materially different provision inserted later.

For anyone designing a reform process, the question is therefore not simply whether public meetings were once held. The record should show which version of the proposal participants saw, what information they received, whether later changes were material, and whether an affected public had a meaningful chance to engage with the substance ultimately advanced. If a proposal changes significantly after consultation, the safest constitutional course may be renewed participation on the new material rather than assuming that earlier engagement covers it.

Not every issue was ripe for a final constitutional answer

Ndii is also a useful lesson in judicial restraint. The Court held that the question whether separate constitutional amendments must always appear as separate referendum questions was not ripe because the IEBC had not yet made the decision that would frame the referendum question or questions. A court ordinarily decides an actual legal controversy, not every dispute that might arise later in the constitutional process. For litigants, that means timing a challenge so that the responsible institution has acted or clearly threatens to act, unless a recognised exception justifies earlier intervention.

The Court also held that the IEBC had sufficient composition and quorum for the signature-verification function at issue, and it did not impose on the IEBC a separate duty under Article 257(4) to ensure that promoters had conducted public participation before verification. Those holdings matter because they stop the case from being rewritten as a judgment that placed every constitutional duty on one institution. Ndii is fundamentally about keeping each actor within the constitutional responsibility allocated to it.

A practical compliance test for a future popular initiative

A future promoter should be able to demonstrate that the initiative genuinely originated with identifiable citizen promoters rather than with an office exercising state power; that the supporting signatures were gathered for the proposal actually being advanced; that the IEBC was allowed to perform its verification function independently; that county assemblies considered the same lawful proposal; that material additions were not smuggled into the process without renewed constitutional scrutiny; and that referendum questions, when required, were framed at the legally appropriate stage. The point is not bureaucratic tidiness. It is democratic ownership and institutional legitimacy.

For advocates challenging or defending an amendment process, the first discipline is to identify the precise proposition for which Ndii is cited. A passage about constituent power may come from an individual opinion without commanding the majority. A challenge to a presidentially initiated Article 257 process stands on firmer binding ground. A public-participation challenge to a late material addition can rely on the unanimous treatment of the Second Schedule. A basic-structure argument must confront the Supreme Court’s rejection of the four-stage doctrine as a binding additional pathway.

Why Ndii matters to ordinary constitutional politics

Constitutions need a way to change, but the method of change can be as important as the content. Article 257 gives citizens an extraordinary route to place constitutional reform on the national agenda without depending at the outset on the executive or Parliament. If state institutions could originate that route with public resources and then present the project as a citizens’ initiative, the constitutional distinction between delegated government power and direct popular initiative would become difficult to defend.

Ndii preserves that distinction while leaving Kenya’s formal amendment architecture intact. Its central message is practical: constitutional reform is not lawful merely because a proposal is popular, politically important or eventually capable of winning votes. The origin, sequence, participation and institutional roles must also comply with the Constitution. Public power cannot cure a defective route by giving it the right name.

Source note

This article was prepared from Attorney General & 2 others v Ndii & 79 others; Dixon & 7 others (Amicus Curiae) (Petitions 12, 11 & 13 of 2021 Consolidated) [2022] KESC 8 (KLR) (31 March 2022); Ndii & others v Attorney General & others (consolidated petitions) [2021] KEHC 9746 (KLR); the Constitution of Kenya 2010, especially Articles 89, 143, 255, 256 and 257; and the Supreme Court’s separate opinions and formal orders in the BBI appeal.

Suggested citation

Ronald Serwanga, “Attorney General v Ndii: Popular Initiative Limits” East Africa Legal Insight (7 September 2026).

Legal information note: This article is for general legal education and does not constitute legal advice for a specific dispute.