The Court of Appeal has upheld Kenya’s Affordable Housing Act in full. It also found, in the same judgment, that the law leaves out the poorest Kenyans it was written for, and warned that its own governing board could become a tool for political patronage. It let the law stand anyway.

Joshua Kimani · Munyeri Advocates · 25 September 2026 · 15 min read

Dr Magare Gikenyi & 5 others and Peter Okiro v The Cabinet Secretary for Lands, Public Works, Housing and Urban Development & 65 others, Civil Appeal No. E832 of 2024 (consolidated with Civil Appeal No. E019 of 2025), Court of Appeal at Nairobi, before Karanja, Kiage, Muchelule, Korir and Odunga, JJA, delivered 25 September 2026.

§1  A law upheld, with reservations attached

Five judges of the Court of Appeal, sitting as Karanja, Kiage, Muchelule, Korir and Odunga, JJA, today dismissed the consolidated appeals against the Affordable Housing Act, 2024, the statute that imposes the 1.5 percent housing levy on every employee’s gross salary and every other Kenyan’s gross income. The appeal came from two individual camps, Dr Magare Gikenyi and four others acting in person, and a sixth appellant represented by counsel, challenging a High Court judgment delivered on 22 October 2024 that had already dismissed six consolidated petitions against the Act.

The case’s own respondent list is worth pausing on. Sixty-six parties are named, including the Kenya Human Rights Commission, Katiba Institute, the Law Society of Kenya, Kituo Cha Sheria, and a run of professional and workers’ unions from medical practitioners to teachers, all of whom had petitioned against the Act at the High Court alongside the appellants. None of them pursued this appeal. Whatever their reasons, resources, tactics, or a considered decision that the High Court’s judgment was not worth contesting further, the fight that reached the Court of Appeal was carried by individual litigants, not Kenya’s most established civil society institutions.

The appellants raised thirty-one grounds of appeal, ranging from the technical, devolution, vagueness, retrospectivity, to the rhetorical, that the Act “entrenches communist ideologies” and subjects Kenyans to “slavery and servitude.” One ground alleged an undisclosed political arrangement between the Presidency and the Judiciary over the housing programme. The Court’s judgment does not engage that allegation anywhere in its reasoning. It is answered, in effect, by silence, a court declining to dignify an unsubstantiated claim with a response rather than issuing a formal rebuke. The substantive grounds fared better, in the sense that the Court took each seriously. It rejected nearly all of them. But rejection, on this record, was not the same as vindication.

§2  The admission buried in the reasoning

The appellants argued that the Act does not deliver affordable housing so much as it compels everyone to pay for a scheme only some will benefit from, a system, as they put it, in which the “haves” buy the houses while the “have-nots” remain homeless despite being taxed to build them. The Court did not simply dismiss this. It examined the Affordable Housing Regulations, 2025, the eligibility rules that govern who actually gets a unit, and reached a striking finding of its own.

Eligibility under the Regulations requires a formal application, a deposit of up to five percent of the purchase price, and the financial capacity to meet ongoing payment obligations the Board must verify before allocation. The Court concluded plainly that this “program, as envisioned under the Act, is meant for those intending to purchase or develop affordable housing. In other words, the low-income earners and the developers. It does not cater for the no income earners,” people the Court itself called “most deserving of the constitutional intervention.” It went further, warning of what happens without a mechanism to identify genuinely deserving beneficiaries and monitor their status over time:

That is an unusually candid line for an appellate bench to commit to a published judgment, five judges naming, in plain language, the exact failure mode critics of the housing levy have argued about since before its enactment. The Court then concluded, at paragraph 82, that “the Act, as enacted, cannot be said to have fully met the constitutional expectation in Article 43(1)(b) of the Constitution.” Having found both the design flaw and the risk it invites, the Court did not strike the Act down, suspend it, or order Parliament to fix the eligibility framework. It reached instead for the doctrine of progressive realisation, the principle that socio-economic rights may be delivered incrementally rather than all at once, and concluded that “more needs to be and must be done going forward,” leaving the “going forward” entirely to the same political process whose product it had just described as vulnerable to capture by cronies.

Progressive realisation is a legitimate doctrine, and courts everywhere are rightly cautious about ordering legislatures to redesign social programmes from the bench. But there is a difference between deferring on how quickly a right is realised and deferring on whether a program the Court itself has found excludes the intended beneficiaries will ever be corrected. The first is judicial restraint. The second risks becoming judicial acquiescence to a defect the judgment itself put on the public record, with no enforcement mechanism attached to the finding.

§3  A violation nobody can act on

A quieter but equally telling moment concerns the Commission on Revenue Allocation. Article 205(1) of the Constitution requires that the Commission be given an opportunity to consider any Bill with financial implications for county governments before it is passed. The High Court had held that the Commission’s input only became necessary once the affordable housing programme reached its implementation stage, after the Bill was already law. The Court of Appeal rejected that reading outright, and did so in unusually direct terms: the High Court, it said, “seemed to have imported an interpretation that went contrary to the express stipulations in Article 205 of the Constitution,” which requires the Commission’s input “before the Bill is discussed in Parliament and not at the stage of its implementation.”

That is, in substance, a finding that the Bill’s passage did not follow the constitutionally required process. It is the same species of defect that sank the Community Groups Registration Act in a High Court ruling delivered fifteen days earlier: a procedural step the Constitution requires before enactment, skipped. Here, though, the finding went nowhere. The Court explained that “the CRA was never made a party to the proceedings and hence its position as to whether or not this opportunity was afforded was not clearly brought out by the petitioners whose duty it was to prove this assertion.”

In other words, the Court identified what it itself called a misreading of an express constitutional requirement, and then declined to act on it because the institution whose exclusion mattered most, the Commission itself, never sued anyone. This is a defensible application of ordinary rules of standing and evidence. It is also a striking illustration of how a genuine constitutional violation can be correctly diagnosed by a court and still go entirely unremedied, not because the violation was not real, but because no party before the court had the standing and the evidence to press it home. A future challenge, properly framed with the Commission joined as a party, would have a considerably stronger foundation than this appeal did.

§4  A promise held on trust

The appellants also argued that building affordable housing on public land and then selling the completed units to private buyers bypasses the National Land Commission, the body the Constitution designates to hold and administer public land in trust for Kenyans. The Court agreed that the Commission’s role here is significant, tracing Article 62 of the Constitution and section 12 of the Land Act to conclude that the Commission’s role in allocating public land for the housing programme “is paramount,” and that “any allocation of public land without the approval of the NLC will, accordingly, be unlawful.”

Having said that, the Court declined to find that any unlawful allocation had actually occurred, calling such a finding “speculative” on the record before it. That leaves the safeguard entirely prospective: a warning that a future allocation without the Commission’s sign-off would be unlawful, rather than a structural check built into how the programme actually operates today. Communities near a specific housing project who believe public land was allocated without the Commission’s involvement will have to litigate that question site by site, after the fact, rather than relying on anything this judgment puts in place in advance.

§5  Two rulings, fifteen days apart

On 10 September 2026, the High Court struck down the Community Groups Registration Act because the record showed submissions had come “from state actors and one quasi-state actor,” with no evidence any ordinary member of the public or community group had been reached, a process the Court called illusory in an often-quoted line: effective public participation, it said, “is not a newspaper advertisement.”

Fifteen days later, on the Affordable Housing Act, the Court of Appeal reached the opposite outcome on the same doctrinal question. The record here showed newspaper advertisements inviting submissions, committee hearings at Parliament Buildings, and public hearings in 19 of Kenya’s 47 counties, leaving 28 counties without a hearing of their own. The appellants also cited a survey, attributed to Trends and Insight Africa, claiming seven in ten Kenyans rejected the housing levy. The Court found the newspaper notices, the hearings, and the parliamentary committee process sufficient, applying a standard of reasonableness that defers heavily to what Parliament itself judged appropriate, and declined to weigh the cited survey at all: “we are unable to determine how authoritative this research is and how the entity that conducted the research arrived at its findings.”

The two cases are not identical, and a careful reader should resist the temptation to call this a straightforward contradiction. The CGRA record showed, on the Court’s own finding, no participation at all by the public the law was meant to regulate. The housing levy record showed real, if geographically incomplete, engagement. Different facts can fairly produce different outcomes under the same legal test. What the pairing does expose is something more structural: Kenya’s “meaningful public participation” doctrine, built case by case since Robert Gakuru in 2014 through the Supreme Court’s BAT decision, has no fixed quantitative floor, no minimum number of counties, no required response rate, no defined channel beyond “reasonable” ones. Two benches applying the same body of authority, fifteen days apart, produced a demanding standard on one file and a highly deferential one on the next, each internally coherent, together telling Parliament almost nothing durable about what the doctrine actually requires the next time.

There is also an asymmetry worth naming plainly. The Court treated an unauthenticated survey suggesting mass public rejection of the levy as too uncertain to weigh, a fair evidentiary instinct. It treated the admitted fact that 28 of 47 counties received no hearing at all with considerably less scrutiny, accepting the remaining 19 as sufficient without asking what share of the population, or which counties, that left out. Caution about weak evidence should cut both ways.

§6  A process checked, a politics untouched

An independent judiciary is cardinal to defending the rule of law and taming government excess in a democratic nation, and this judgment shows that function working exactly as it should in places: a court testing a tax-and-spend programme against the text of the Constitution, finding real defects, and saying so on the record for everyone to see. No nation on earth, though, has ever been salvaged from collapse by its judiciary alone, and this judgment is itself an argument for why. Confronted with a survey claiming seven in ten Kenyans opposed the very levy it was asked to review, the Court did not treat that opposition as decisive, and it should not have. As the South African Constitutional Court put it in a passage this bench itself relied on, “there is no authority for the proposition that the views expressed by the public are binding on the legislature if they are in direct conflict with the policies of Government.” Public participation, on that view, exists to inform and supplement democratic government, “not to conflict with or even overrule or veto” what elections and majority rule produce.

That is, in the Court’s own words, the honest limit of what litigation against an unpopular law can achieve. A court can find that a tax scheme’s own eligibility rules exclude the poorest Kenyans it claims to serve. It can find that a constitutionally required consultation was skipped. It can warn, unusually bluntly, that a housing board risks becoming a patronage machine. What it cannot do, and did not purport to do here, is force Parliament to rewrite the eligibility rules, compel the executive to widen the consultation next time, or make the housing levy popular. A court can strike down a bad process. It cannot compel a good politics. Revolutions, if they come, do not happen in courts of law. They happen at the ballot box, and in the street, where a public that already showed once, in the Gen Z-led protests of 2024, that it will make its displeasure with a tax measure impossible to ignore, decides whether the next Finance Bill, the next housing regulation, or the next election manifesto answers what this judgment left standing.

Gerald Rosenberg’s study of American courts, The Hollow Hope: Can Courts Bring About Social Change?, is worth returning to here for the same reason it mattered in our reading of the CGRA ruling: judicial victories, and judicial findings that fall short of victories, tend to be hollow when they are not backed by sustained political pressure that makes the underlying conduct costly to repeat. A housing board that this Court itself flagged as vulnerable to cronyism will only avoid that fate if someone outside the courtroom, voters, county governments, the press, keeps watching how it allocates units once the judgment is a week old and the headlines have moved on.

§7  What would make this ruling mean something

The judgment itself supplies the agenda, if anyone takes it up:

Bottom line

The Affordable Housing Act survives this appeal intact, and the judgment that saved it reads, in long stretches, like a list of reasons it should not have. A five-judge bench found a design that locks out the neediest, warned in unusually blunt language of a road to political patronage, and identified a skipped constitutional consultation, then upheld the law on each point regardless. None of that makes the judgment wrong on the law it was asked to apply. It does mean the questions this ruling raises about who the housing levy actually serves were not answered here. They will be answered, if they are answered at all, the way this newspaper’s earlier analysis of the CGRA ruling argued they usually are: not by a further judgment, but by whether the public that already forced one government climbdown in 2024 keeps paying attention to what a board this Court itself has warned about does next.

Sources and further reading: Dr Magare Gikenyi & 5 others and Peter Okiro v The Cabinet Secretary for Lands, Public Works, Housing and Urban Development & 65 others, Civil Appeal No. E832 of 2024 (consolidated with Civil Appeal No. E019 of 2025), Court of Appeal at Nairobi (25 September 2026); the Affordable Housing Act, 2024 via Kenya Law; Mitu-Bell Welfare Society v Kenya Airports Authority [2021] KESC 34; Government of the Republic of South Africa v Grootboom [2000] ZACC 19; Merafong Demarcation Forum v President of the Republic of South Africa [2008] ZACC 10; and Gerald N. Rosenberg, The Hollow Hope: Can Courts Bring About Social Change? (University of Chicago Press). Our earlier analysis of the Community Groups Registration Act ruling, referenced in Section 5 and Section 6 above, is available elsewhere on this site.

Joshua Kimani Joshua Kimani is a lawyer, constitutional law scholar and writer at Munyeri Advocates, where his practice and scholarship focus on public law, devolution and civic space.

This article is provided for general information only and does not constitute legal advice. It reflects the author’s own analysis and opinion on a live matter of public interest law, and readers with a specific legal question should seek advice on their particular facts.