KE-A-03: The Second Liberation — The Return of Multiparty Democracy (1990–2002)

1. Key Takeaways

• The "Second Liberation" is the name Kenyan pro-democracy activists gave to the 1990–2002 struggle, deliberately framing it as the unfinished business of the first liberation of 1963. The phrase asserted that uhuru (independence) from colonial rule had delivered a flag and an anthem but not popular sovereignty, and that the one-party KANU state under Daniel arap Moi was a domestic continuation of the centralising, repressive late-colonial Provincial Administration architecture (KE-A-01). The framing was politically consequential: it allowed lawyers, clergy, detainees, and exiled academics to claim the nationalist mantle for the reform cause rather than ceding it to KANU, and it located the demand for multipartism within a longer Kenyan emancipatory narrative rather than presenting it as a foreign import. The term recurs across the memoir and movement literature — in the writings of Willy Mutunga, in the rhetoric of the Law Society of Kenya, and in the church pronouncements of the period — and is the organising concept of this document.

• By the late 1980s the KANU one-party state had concentrated power to a degree that left no lawful avenue for opposition, making constitutional reform itself the central political demand. The 9 June 1982 insertion of Section 2A had made KANU "the sole political party in Kenya" (KE-A-02); the 1986 abolition of security of tenure for the Attorney General and Auditor General and the 1988 extension of that abolition to judges had gutted institutional checks; and the introduction of mlolongo — the queue-voting system — for the 1988 KANU primaries had destroyed even the secrecy of the ballot. With Parliament, judiciary, civil service, and electoral machinery all subordinated to the Presidency, reformers concluded that no incremental change was possible without first repealing Section 2A and rewriting the constitution. This is why the Second Liberation was, from the outset, a katiba (constitution) struggle as much as a party-pluralism struggle.

• The 1988 mlolongo election was the nadir that radicalised the reform constituency. In the KANU primaries of that year, voters were required to queue publicly behind a photograph or representative of their preferred candidate, and the official tallies were widely manipulated — candidates with visibly shorter queues were declared winners in several constituencies. The most-cited instance is the Kiharu result in which the official count contradicted the visible queue [TBD-VERIFY: precise Kiharu queue figures and the specific contested constituencies vary across sources]. The system's transparent fraudulence discredited the one-party order in the eyes of many who had previously accommodated it, and it directly motivated the 1990 decision by Kenneth Matiba — himself a KANU insider until 1988 — and Charles Rubia to demand multipartism.

• The murder of Foreign Minister Robert Ouko on 13 February 1990 and the Saba Saba rally of 7 July 1990 converted elite reform sentiment into mass mobilisation. Ouko's killing (KE-A-02) broke the assumption that the Moi system would not destroy its own senior figures and energised the cumulative reform demand. When Matiba and Rubia called a rally at Kamukunji grounds for 7 July 1990 to demand the repeal of Section 2A, the state banned the gathering and detained the organisers on 4 July; the rally proceeded informally and the heavy-handed security response produced deaths and a several-day disturbance across Nairobi and Central Province. Saba saba — "seven seven" — became the movement's permanent calendar anchor, re-invoked in 1997 and beyond. [TBD-VERIFY: the Saba Saba death toll is variously given as ~20 (official) to 39 or more (Africa Watch); no post-2003 commission established a definitive figure.]

• The repeal of Section 2A on 3 December 1991 was overdetermined, and the corpus does not adjudicate a single cause. Three readings persist (detailed in Section 7). The domestic-heroism reading credits the Saba Saba mobilisation, the church (Bishops Gitari, Okullu, the martyred Muge), the Law Society, the detainees, and the exiled critics. The donor-conditionality reading credits the November 1991 Paris Consultative Group's suspension of roughly US$350 million in quick-disbursing aid [TBD-VERIFY exact figure] under post-Cold-War governance conditioning, championed by US Ambassador Smith Hempstone. The elite-pact / Moi-strategic-calculation reading argues Moi conceded on his own terms because he correctly foresaw that a fragmented opposition would lose under plurality rules. All three operated simultaneously.

• The opposition's fragmentation was the structural condition that re-elected Moi in both 1992 and 1997. The Forum for the Restoration of Democracy (FORD), registered as the umbrella opposition in 1991, split into FORD-Asili (Matiba) and FORD-Kenya (Oginga Odinga) through 1992, with Mwai Kibaki forming the Democratic Party. Moi won the 29 December 1992 election with 36.35 percent against an aggregate opposition majority [TBD-VERIFY tallies], and repeated the result in 1997 with 40.40 percent (KE-A-02). Kenya's first-past-the-post rule and the requirement to win 25 percent in five of eight provinces converted opposition disunity into incumbent victory — the precise outcome the strategic-calculation reading of the 1991 repeal had anticipated.

• The Rift Valley "land clashes" of 1991–93 and the 1997 Coast violence are among the most contested episodes of the era, turning on whether the violence was spontaneous or state-orchestrated. Human Rights Watch's Divide and Rule (1993) and the National Council of Churches' The Cursed Arrow (1992) documented attacks on Kikuyu, Luhya, and Luo-Kisii communities in Kalenjin-majority areas, timed to the multiparty transition. The 1999 Akiwumi Report identified senior KANU figures as having allegedly incited or facilitated the violence but produced no prosecutions. The spontaneous reading treats the clashes as majimbo-era ethnic grievance erupting under electoral competition; the orchestrated reading treats them as a deliberate KANU strategy to depopulate opposition-voting areas and validate the warning that multipartism would bring chaos. [TBD-VERIFY: cumulative 1991–93 casualty figures of ~1,500 dead and ~300,000 displaced are widely cited but contested.]

• The Goldenberg scandal was the era's defining corruption episode and a measure of the patronage-state under fiscal stress. Goldenberg International, controlled by Kamlesh Pattni, was paid export-compensation for largely fictitious gold and diamond exports; the 2005 Bosire Commission estimated the loss at roughly KSh 158.3 billion [TBD-VERIFY: estimates range widely]. Operating through the structural-adjustment decade, Goldenberg illustrated how the one-party patronage system adapted to donor-imposed austerity by capturing the central bank's foreign-exchange machinery — and it became a central exhibit in the reform movement's case against the Moi state.

• The constitutional-review movement of 1997–2002 split into rival processes — the parliamentary CKRC and the civil-society Ufungamano Initiative — before merging. The IPPG reforms of 1997 produced the Constitution of Kenya Review Act, but civil society rejected the resulting Constitution of Kenya Review Commission (CKRC) as government-controlled and convened the parallel Ufungamano Initiative under religious leadership. The two tracks merged in 2001 under Yash Pal Ghai's chairmanship; the process Ghai led produced draft constitutions that, after the 2005 referendum defeat (KE-B-02, when written) and the 2007–08 crisis, ultimately yielded the 2010 Constitution (KE-A-04). The Second Liberation thus did not complete its katiba project in its own decade.

• The 27 December 2002 NARC victory ended KANU's 39-year rule and is the Second Liberation's culminating event, though contested in its meaning. Moi's October 2002 imposition of Uhuru Kenyatta as KANU's candidate triggered the defection of senior KANU figures into the Raila Odinga–led Rainbow Alliance, which fused with Kibaki's National Alliance to form the National Rainbow Coalition (NARC). Kibaki defeated Uhuru 62.2 percent to 31.3 percent and Moi conceded promptly (KE-A-02, KE-B-01). Whether 2002 represented the consummation of the activist Second Liberation or merely an elite alternation that left the patronage-state intact is the central interpretive question carried forward into the Kibaki, Kenyatta, and Ruto eras.

---

2. The Closed System — The KANU One-Party State on the Eve of the Struggle

To understand the Second Liberation it is necessary to grasp the completeness of the political closure that confronted reformers at the end of the 1980s. The parent era-document (KE-A-02) traces the institutional consolidation in detail; this section establishes the specific configuration of repression and exclusion that made constitutional reform — rather than mere electoral competition — the indispensable first demand.

The legal foundation was Section 2A of the Constitution, inserted on 9 June 1982, which declared KANU "the sole political party in Kenya." The amendment had been moved in a single parliamentary afternoon following the attempt by Oginga Odinga and George Anyona to register a rival party, and it converted into constitutional law the de facto single-party order that had operated since the 1969 proscription of the Kenya People's Union (KE-A-01). The consequence was that the formation of, or membership in, any party other than KANU was a constitutional impossibility; the only contestation permitted was within KANU, under rules the party controlled.

That internal contestation was itself progressively hollowed. Three amendments between 1986 and 1988 removed the institutional checks that might have constrained executive power. The 1986 constitutional amendment abolished security of tenure for the Attorney General, the Auditor and Comptroller General, converting these watchdog offices into instruments removable at presidential discretion. The 1988 amendment extended the abolition to High Court and Court of Appeal judges, producing a bench that, through the late 1980s, consistently ruled for the executive in politically charged matters. The reform movement's later insistence on an independent judiciary, an autonomous prosecutorial office, and constitutionally entrenched tenure — all of which appear in the 2010 Constitution (KE-A-04) — was a direct response to this specific dismantling.

The security apparatus operated through detention without trial under the Preservation of Public Security Act and through the interrogation chambers in the basement of Nyayo House on Kenyatta Avenue, where, as documented in post-2003 Truth Justice and Reconciliation Commission proceedings and in survivor testimony, torture practices including water-immersion, sleep deprivation, and prolonged solitary confinement were applied. The 1986–88 "Mwakenya" affair — directed at an alleged clandestine network linked to the exiled writer Ngugi wa Thiong'o — produced the detention and prosecution of approximately seventy-five individuals [TBD-VERIFY: figures vary across sources], many held in the Nyayo House cells. Raila Odinga, son of Oginga Odinga, was detained on three separate occasions across the decade for cumulative periods exceeding six years (KE-A-02), an experience that forged his identity as the foremost opposition-mobilisation figure of the 1990s.

The economic context compounded the political closure. The 1990s were Kenya's structural-adjustment decade: IMF and World Bank programmes conditioned macroeconomic support on fiscal restraint, parastatal reform, and trade liberalisation, while GDP per capita stagnated or declined in real terms [TBD-VERIFY specific figures] and unemployment among educated youth rose. The patronage system that bound KANU's ethnic-regional coalition together continued to require fiscal flexibility even as austerity constrained it — a tension that drove both elite defection and the corruption episodes, above all Goldenberg, that the reform movement would weaponise. By 1989, then, the reformers faced a state that had foreclosed lawful opposition, captured its own oversight institutions, retained a functioning detention-and-torture apparatus, and presided over an economy under acute external stress. The only lever capable of opening this system was constitutional change, and that recognition is the starting point of the Second Liberation.

---

3. Mlolongo and the Discrediting of the One-Party Order

The KANU general election of 1988 supplied the reform movement with its founding grievance. In the party primaries of that year the Moi administration introduced mlolongo — the queue-voting system — under which voters lined up publicly behind the photograph, the agent, or a representation of their preferred candidate, and the queues were counted by KANU officials. A candidate who secured 70 percent or more of the queue was declared elected outright without a secret-ballot stage. The system was defended publicly as a Kenyan innovation expressing communal openness; in practice it stripped voters of the secret ballot's protection against retaliation by employers, chiefs, and local KANU bosses, and it placed the count entirely in the hands of party functionaries.

The fraud was visible. In several constituencies, candidates with manifestly longer queues were declared losers, and the official results contradicted what assembled crowds had seen with their own eyes. The most-cited instance involved a Murang'a-district contest in which the announced outcome reversed the visible queue, an episode that contributed directly to Kenneth Matiba's alienation from KANU [TBD-VERIFY: the specific constituency, candidates, and queue figures are reported variably across the secondary literature]. Matiba, a wealthy Murang'a businessman and Cabinet Minister who had been a KANU insider, resigned from the Cabinet and Parliament in December 1988 in protest at the manipulation — a resignation that would prove pivotal eighteen months later.

The 1988 general election compounded the primary fraud. Because Section 2A made KANU the only lawful party, the parliamentary "election" was in effect the conclusion of the KANU nomination process, and the mlolongo primaries determined who would sit in Parliament. Candidates who crossed the 70-percent threshold in the queue stage were declared elected without any secret-ballot run-off, which gave KANU returning officers and provincial administrators decisive discretion over which queues to count and how. The system also exposed civil servants, teachers, and employees of parastatals to direct observation of their political choices, with predictable consequences for those who queued behind disfavoured candidates. Contemporary accounts in the Weekly Review and the documentation later assembled in Africa Watch's Kenya: Taking Liberties (1991) recorded widespread complaints of rigging, intimidation, and the reversal of visible results — a body of evidence that the reform movement would marshal as proof that the one-party order had abandoned even the pretence of fair internal competition.

The episode also illuminated the wider pattern of institutional capture documented in Section 2. With the judiciary stripped of tenure protection in the 1988 amendment passed in the same period, aggrieved candidates had no realistic recourse to the courts; with Parliament an extension of the executive, no legislative remedy was available; and with the press operating under licensing and sedition constraints, even documentation was hazardous. Mlolongo was thus not an isolated electoral malpractice but the visible surface of a comprehensively closed system, and its very brazenness — conducted in the open, before assembled crowds — made it uniquely difficult for the regime to deny.

The mlolongo episode mattered for three reasons. First, it discredited the one-party order among constituencies that had previously accommodated it: the system's transparent dishonesty made it impossible to defend the claim that KANU's internal competition substituted adequately for multiparty choice. Second, it widened the reform constituency beyond the intellectuals and detainees who had borne the brunt of 1980s repression to include disaffected KANU insiders such as Matiba — a broadening that gave the 1990 mobilisation its cross-elite reach. Third, it crystallised the analytical conclusion that the problem was structural rather than personnel-based: no honest result was possible within a system whose rules and counting machinery the ruling party controlled. The 1988 election thus functioned as the nadir from which the Second Liberation took its energy, and its memory recurs throughout the movement's later rhetoric as the definitive proof that the one-party state could not reform itself from within.

---

4. The Church and Civil Society — The Opening of Political Space

In a system that had detained politicians, captured the judiciary, and silenced Parliament, the institutions with the standing to challenge the state were those it could least easily suppress: the churches and the legal profession. Their mobilisation between 1987 and 1991 created the political space within which open opposition became possible, and it is the empirical core of the "domestic-heroism" reading of the Second Liberation.

The Anglican and Presbyterian churches, together with the Catholic episcopate and the National Council of Churches of Kenya (NCCK), provided the era's most consistent institutional voice for reform. Three Anglican bishops were especially prominent. Henry Okullu, Bishop of Maseno South, used sermons and pastoral letters from the mid-1980s to attack detention without trial, the abolition of judicial tenure, and the mlolongo system, and to argue for the restoration of multiparty competition. David Gitari, Bishop of Mount Kenya East (and later Archbishop), preached against electoral manipulation and state violence and survived what his supporters characterised as an assassination attempt at his Kerugoya home in 1989 [TBD-VERIFY: the circumstances of the 1989 incident at Gitari's residence are contested]. Alexander Muge, Bishop of Eldoret, was the most confrontational; he died on 14 August 1990 in a road collision near Eldoret shortly after a Cabinet minister had publicly warned him not to set foot in Busia — a sequence that produced enduring suspicion of foul play, though no inquiry established culpability (KE-A-02). [TBD-VERIFY: the Muge death remains officially a road accident; the warning-and-death sequence is documented but causation is unproven.]

The clergy's intervention was theologically framed and therefore difficult for the state to delegitimise as foreign-sponsored subversion. By grounding the reform demand in pastoral concern for the persecuted and in a Christian account of just government, the bishops claimed a moral authority that detained politicians, tarred as tribalists or agents of foreign powers, could not. The Catholic bishops' pastoral letters and the NCCK's documentation work — culminating in the 1992 report The Cursed Arrow on the ethnic clashes — supplied both moral cover and an evidentiary record that the international community could cite.

The legal profession supplied the technical and litigational dimension. The Law Society of Kenya (LSK), under reform-minded leadership including Paul Muite (who became LSK chairman in 1991), Kiraitu Murungi, Willy Mutunga, Gibson Kamau Kuria, and Pheroze Nowrojee, challenged detention orders, defended political detainees, and articulated the constitutional case for repeal of Section 2A. Gibson Kamau Kuria was himself detained in 1987 after filing habeas-corpus applications on behalf of detainees. Willy Mutunga's later account, Constitution-Making from the Middle (1999), documents how this lawyer-activist network theorised the reform demand as a constitutional project — locating it in the language of rights and process rather than in the language of party competition alone, and thereby shaping the katiba movement that would run through the rest of the decade and into the 2010 Constitution.

The church and bar were effective precisely because they occupied institutional ground the one-party state could not easily seize. KANU could detain a politician, expel an MP, or rig a queue, but it could not credibly proscribe the national churches, whose congregations crossed every ethnic and regional line and whose international ecumenical connections amplified domestic repression abroad. Nor could it easily silence the bar without abandoning the legal forms by which it claimed legitimacy. The reformers exploited this asymmetry deliberately: clergy delivered the demands that detained politicians could not voice from prison, and lawyers translated those demands into constitutional and rights-based claims that the regime found harder to dismiss as tribal agitation. Widner's The Rise of a Party-State in Kenya (1992) and Mutunga's Constitution-Making from the Middle (1999) both stress this institutional-niche dynamic as central to explaining how an apparently total system was prised open from within.

A third strand was the returning and resident intelligentsia and the émigré opposition. Exiled figures such as the academics and former detainees associated with the London-based opposition kept Kenyan repression visible to international audiences, while resident academics, journalists at Hilary Ng'weno's Weekly Review and the Nation, and human-rights monitors built the domestic documentary record. The cumulative effect of church, bar, press, and intelligentsia was to manufacture a political space the one-party state could neither fully occupy nor entirely suppress — and it was into that space that Matiba and Rubia stepped in 1990.

---

5. Saba Saba — 7 July 1990

The Saba Saba rally is the symbolic fulcrum of the Second Liberation, the moment at which elite reform sentiment became mass political claim. In late May and June 1990, Kenneth Matiba and Charles Rubia — the former a resigned Cabinet minister, the latter a former Mayor of Nairobi — publicly demanded the repeal of Section 2A and announced a rally at Kamukunji grounds in Nairobi, the historic nationalist meeting-ground, for 7 July 1990. They were soon joined in the public demand by the veteran Oginga Odinga and by the lawyer-activists of the LSK.

The state's response was repression. The rally was banned. On 4 July 1990, Matiba and Rubia were detained without trial, along with the lawyer Raila Odinga and others; the detentions were intended to behead the mobilisation before it could gather. They did not. On 7 July, crowds assembled across Nairobi and spread to parts of Central Province and the Rift Valley town of Nakuru; the security forces — the General Service Unit and regular police — dispersed gatherings with force, and the resulting several-day disturbance produced deaths, injuries, mass arrests, and property destruction. The violence was concentrated in Nairobi's Eastlands and along the Nairobi–Nakuru corridor.

The human cost is contested. Official figures gave approximately twenty deaths over the 7–10 July period; Africa Watch and other contemporary monitors gave figures from around thirty-nine to over a hundred [TBD-VERIFY: no post-2003 commission established a definitive Saba Saba death toll]. Several hundred people were arrested, and a number of activists and bystanders were charged. The disparity between official and independent counts itself became a movement grievance, evidence of a state that could neither acknowledge nor account for the violence it deployed.

Saba Saba's significance lay less in its immediate outcome — Section 2A was not repealed for another seventeen months — than in what it demonstrated. First, it showed that the Moi administration had lost the capacity to suppress reform sentiment through selective detention alone: removing the organisers did not prevent the mobilisation. Second, it converted the reform demand from an elite-intellectual claim into a mass political event with its own calendar anchor; "Saba Saba" entered Kenyan political vocabulary permanently and was re-invoked in the "Saba Saba II" protests of 1997 and in subsequent decades. Third, it internationalised the Kenyan reform question at precisely the moment, in 1990, when the end of the Cold War was loosening Western tolerance of allied authoritarianisms — making Saba Saba both a domestic turning point and the event that connected the domestic struggle to the donor pressure that would crystallise in 1991. The interplay of those two forces is the subject of the analysis that follows.

---

6. The Repeal of Section 2A — December 1991

The seventeen months between Saba Saba and the repeal of Section 2A saw the convergence of escalating domestic pressure and a transformed international environment. Through 1991, the reform movement institutionalised. In August 1991 a group of opposition figures — Oginga Odinga, Masinde Muliro, Martin Shikuku, Ahmed Bamahriz, George Nthenge, and Salim Ndamwe (the "Young Turks" and veterans together) — announced the formation of the Forum for the Restoration of Democracy (FORD), initially as a pressure group, since registering it as a party was constitutionally barred. FORD's deliberate self-presentation as a broad national front, spanning Luo, Luhya, Kikuyu, and coastal constituencies, was an attempt to pre-empt the ethnic fragmentation that everyone understood to be the opposition's central vulnerability. The state harassed FORD's rallies and briefly re-detained Odinga, but the front's existence demonstrated that the one-party prohibition was becoming unenforceable in practice.

The international dimension transformed in the second half of 1991. The collapse of the Soviet bloc removed Kenya's Cold-War value to Western governments as an anti-communist anchor in a volatile region, and it emboldened Western donors to attach governance conditions to aid that they had previously withheld. US Ambassador Smith Hempstone, appointed in 1989, had from his arrival publicly criticised the Moi administration's human-rights record and signalled American sympathy for the reform cause — a stance he recounts, with characteristic self-regard, in his memoir Rogue Ambassador (1997), which must be read as an interested participant-source. The decisive event came at the Paris Consultative Group meeting of donor governments and multilateral lenders on 25–26 November 1991, at which the donors suspended approximately US$350 million in quick-disbursing balance-of-payments support pending governance reforms, including multiparty competition, judicial independence, and anti-corruption measures [TBD-VERIFY: the precise figure and the exact composition of suspended aid vary across Hornsby (2012), Hempstone (1997), and Branch (2011)].

The Paris suspension produced an immediate fiscal crisis for a government already under structural-adjustment stress, and the response was rapid. On 2–3 December 1991, the KANU Special Delegates Conference at the Kasarani sports complex resolved to repeal Section 2A; Moi addressed the delegates at length, framing the return to multipartism as a sovereign Kenyan decision rather than a capitulation to foreign pressure. Parliament formally enacted the repealing amendment on 10 December 1991. Critically, the repeal removed the single-party prohibition but left intact the wider architecture that would shape the multiparty elections to follow: the imperial presidency, the executive's control over the electoral machinery, the absence of any presidential-runoff requirement, and the 25-percent-in-five-of-eight-provinces threshold. The reform movement had won the right to organise; it had not won the constitutional redesign it had also demanded, and that incompleteness set the terms of the decade ahead.

Two features of the repeal deserve emphasis because they shaped everything that followed. The first is its form: the repeal was achieved through the same machinery — a KANU delegates' conference followed by a parliamentary amendment — that had imposed the one-party state in 1982, which meant that the ruling party retained the initiative throughout and conceded only the specific provision under pressure, not the broader system of executive dominance. The reformers had demanded a comprehensive constitutional conference; what they received was a single repealing amendment passed on KANU's terms and timetable. The second feature is the speed of Moi's pivot from defiance to concession. As late as the first half of 1991 Moi and senior KANU figures had publicly insisted that multipartism would import ethnic chaos and that Kenya's one-party system was a superior, African-rooted form of unity; within days of the Paris suspension that position was abandoned. The abruptness of the reversal is the single strongest piece of evidence for the donor-conditionality reading, while the retention of the surrounding architecture is the strongest evidence for the strategic-calculation reading — which is why the causation has remained genuinely contested rather than settled.

The repeal also reset the opposition's strategic problem. So long as Section 2A stood, the reform coalition had a single, unifying objective whose moral clarity held disparate constituencies together. Once the prohibition fell, that unifying objective dissolved into the divisive question of who would lead the now-lawful opposition and on what ethnic-regional base — the question that fractured FORD within months (Section 8). In this sense the repeal was both the Second Liberation's first great victory and the moment at which its central weakness, opposition fragmentation, was activated.

---

7. Three Accounts of the Liberalisation

Why did Kenya liberalise in 1991? The historiography offers three readings, and the corpus does not adjudicate among them, holding that each captures part of the causal architecture.

The domestic-heroism reading — developed most fully by Throup and Hornsby (1998), by Mutunga (1999), and by Widner (1992), and adopted in narrative form by Branch (2011) — credits the cumulative domestic mobilisation: the mlolongo delegitimation of 1988, the Ouko murder, Saba Saba, the church witness of Okullu, Gitari, and the martyred Muge, the LSK's litigation and advocacy, the detainees' sacrifice, and FORD's organisation. On this account the November 1991 donor pressure was a precipitating accelerant of a reversal that domestic dynamics had already made unavoidable; the agency belongs to Kenyans who risked detention, torture, and death. This reading is the one Second Liberation activists themselves advanced, and Makau Mutua's Kenya's Quest for Democracy: Taming Leviathan (2008) gives it a sustained scholarly articulation, treating the reform struggle as a domestically-generated confrontation with an over-mighty state.

The donor-conditionality / external-pressure reading credits the November 1991 Paris suspension and the post-Cold-War transformation of Western policy as the operative cause. On this account, advanced in some policy commentary and implicit in Hempstone's memoir, the domestic movement — for all its courage — lacked the structural leverage to compel a regime that controlled the security forces and the treasury; it was the fiscal shock of the aid freeze, against the new permissibility of governance conditioning after the Soviet collapse, that converted reform demand into reform implementation. The timing is the strongest evidence: the repeal followed the Paris meeting by days.

The elite-pact / intra-KANU-defection / Moi-strategic-calculation reading — associated with Throup and Hornsby's analysis and with Widner — argues that Moi accepted multipartism on his own terms because he had calculated, correctly, that a fragmented and ethnically divided opposition would lose under plurality rules with provincial thresholds. On this account the repeal was less a defeat than a strategic re-positioning: Moi exchanged the legitimacy cost of Section 2A for the political advantage of contested-but-winnable elections, and the subsequent 1992 and 1997 results vindicated the calculation. A related strand emphasises the role of intra-elite manoeuvre — the defection of figures such as Matiba and Kibaki from KANU, and the bargaining among elites over the terms of competition — over mass mobilisation as the proximate driver.

A note on the stakes of the contest is warranted, because the choice among readings is not merely academic. The domestic-heroism reading underwrites a particular national memory — one in which Kenyans liberated themselves and in which the detainees, clergy, and lawyers of the period are the authentic authors of the post-2002 order; it is the reading inscribed in the "Second Liberation" name itself and defended by participants such as Mutunga and Mutua. The donor-conditionality reading, by contrast, implies a more sobering account of African democratisation in which domestic agency was necessary but insufficient and external leverage was decisive — a conclusion with uncomfortable implications for democratic ownership and for the durability of reforms achieved under conditionality. The strategic-calculation reading is the most deflationary of all: it implies that the liberalisation was, at the decisive margin, a regime survival strategy rather than a popular triumph, and that the subsequent decade of KANU electoral victories was not the betrayal of the Second Liberation but its predictable design feature. Each reading thus carries a different verdict on what kind of event 1991 was, and the corpus's refusal to adjudicate is a deliberate preservation of that interpretive openness rather than an evasion.

The corpus position is that the three readings are not mutually exclusive and that the repeal's specific timing and form reflected their interaction. The domestic movement created the legitimacy crisis and the organised constituency; the donor pressure supplied the fiscal trigger and the international permission; and Moi's strategic calculation determined that the concession took the form of plurality multipartism rather than the deeper constitutional redesign the reformers had sought. Each reading, taken alone, is incomplete; together they account for both the fact and the manner of the 1991 reversal.

---

8. The Fragmentation of FORD and the 1992 Election

The opposition's central vulnerability — the one Moi's strategic calculation had identified — materialised almost immediately. FORD, the broad front that had embodied opposition unity in 1991, fractured through 1992 along the fault lines of personality, generation, and ethnicity. The proximate dispute was over the FORD presidential nomination: the veteran Oginga Odinga and the resigned Cabinet minister Kenneth Matiba both claimed the candidacy, and the contest between the "old guard" around Odinga and the "Young Turks" around Matiba and Paul Muite proved irreconcilable. By October 1992 FORD had split into two registered parties: FORD-Asili ("Original"), led by Matiba with Martin Shikuku, drawing on Kikuyu and Luhya support; and FORD-Kenya, led by Oginga Odinga, drawing on Luo and a slice of the reform intelligentsia. Mwai Kibaki, who had resigned from the Moi Cabinet in December 1991, had meanwhile formed the Democratic Party (DP), anchored in Mount Kenya Kikuyu and professional constituencies.

The result was a three-way principal opposition contest that handed Moi the structural advantage. In the 29 December 1992 election — Kenya's first multiparty general election since 1969 — Moi was returned with 1,962,866 votes (36.35 percent), against Matiba's 1,404,266 (26.0 percent), Kibaki's 1,050,617 (19.45 percent), and Odinga's 944,564 (17.48 percent) [TBD-VERIFY: precise tallies vary across the Electoral Commission report, Hornsby (2012), and Throup-Hornsby (1998)]. The aggregate opposition vote of roughly 63 percent exceeded Moi's plurality, but first-past-the-post combined with the 25-percent-in-five-provinces requirement gave Moi the presidency: he cleared the provincial threshold in Rift Valley, Western, Coast, North-Eastern, and Eastern, while each opposition candidate's support was geographically concentrated. The election was monitored by the domestic National Election Monitoring Unit (NEMU), the Commonwealth Observer Group, and others; the Commonwealth group judged that the result broadly reflected the electorate's will despite serious irregularities, while NEMU and most domestic monitors were sharply more critical, documenting systematic registration, transport, and tallying irregularities favouring KANU.

For the Second Liberation, the 1992 result was a sobering lesson. The movement had achieved multipartism but discovered that multipartism alone, under the inherited electoral rules and against a fragmented opposition, reproduced KANU rule. The conclusion many reformers drew — that the unfinished business was constitutional, requiring a redesign of the electoral and executive architecture — became the organising premise of the constitutional-review movement of the later 1990s.

---

9. The 1991–93 Rift Valley Clashes — Spontaneous or Orchestrated?

The 1992 election was conducted against a background of organised ethnic violence in the Rift Valley, Western Province, and parts of the Coast, beginning in late 1991 and continuing into 1993. The violence was concentrated in the Rift Valley areas where Kalenjin and Maasai communities adjoined the Kikuyu, Luhya, and Luo-Kisii populations who had acquired land there through post-independence resettlement — communities that disproportionately supported the opposition. Armed groups, frequently described as Kalenjin "warriors," attacked these settler communities, burning homes and crops and driving people from land; the cumulative toll is estimated at approximately 1,500 dead and around 300,000 displaced for the 1991–93 period [TBD-VERIFY: figures vary significantly across sources; the Akiwumi Report gave district-level figures whose totals depend on aggregation methodology].

The interpretation of this violence is one of the most contested questions of the era, and it turns on whether the clashes were spontaneous or state-orchestrated.

The spontaneous reading treats the clashes as the eruption, under the pressure of electoral competition, of genuine ethnic grievance over land — grievance rooted in the majimbo (regionalist) claims of the KADU tradition (KE-A-01), which held that the Rift Valley was Kalenjin and Maasai ancestral land onto which other communities had been settled. On this account the violence was a horizontal inter-communal conflict that the multiparty transition unleashed rather than one the state manufactured, and KANU's culpability lay at most in rhetorical incitement and failure to protect, not in organisation.

The orchestrated reading — advanced by Human Rights Watch / Africa Watch in Divide and Rule: State-Sponsored Ethnic Violence in Kenya (1993), by the National Council of Churches in The Cursed Arrow (1992), and supported in part by the 1999 Akiwumi Report — treats the clashes as a deliberate KANU strategy. On this account, senior figures incited and financed the attacks to depopulate opposition-voting areas before the election, to punish communities that had embraced multipartism, and to validate KANU's central campaign warning that multipartism would bring majimbo-style ethnic chaos. Divide and Rule documented patterns — coordinated timing, transported attackers, uniform tactics, and security-force passivity or complicity — that it argued were inconsistent with spontaneous communal conflict. The Akiwumi Commission, appointed in 1998 and reporting in 1999, heard extensive evidence and named senior figures including Cabinet ministers Nicholas Biwott and William ole Ntimama as having allegedly incited or facilitated the violence [TBD-VERIFY: the Akiwumi Report's specific findings and named individuals are reported variably; the government rejected parts of the report and no senior figure was prosecuted].

The corpus position notes that the two readings are not wholly exclusive: pre-existing land grievance supplied the combustible material, while the documented patterns of coordination, financing, and security-force complicity support the conclusion that the material was deliberately ignited and directed for electoral advantage. The decisive fact for the Second Liberation narrative is the impunity: no senior figure was ever prosecuted, and the failure of accountability for the 1990s clashes is widely treated as an antecedent of the far larger 2007–08 post-election violence (KE-B-01, KE-B-03), establishing a pattern in which ethnic violence served electoral strategy without legal consequence.

---

10. The Mid-Decade — Goldenberg, Structural Adjustment, and Renewed Mobilisation

Between the 1992 and 1997 elections, two developments shaped the reform struggle: the unfolding of the Goldenberg scandal as the era's defining corruption episode, and the regrouping of civil society around an explicit constitutional-reform agenda.

The Goldenberg scheme, operating from approximately 1990, paid the private company Goldenberg International — controlled by Kamlesh Pattni — export-compensation for purported gold and diamond exports, at a rate of 35 percent on top of export revenues, exceeding the standard 20 percent. The fundamental problem was that Kenya exports neither gold nor diamonds at meaningful scale; the exports were largely fictitious, and the scheme functioned as a mechanism for extracting foreign exchange and Treasury funds. The 2005 Bosire Commission of Inquiry, appointed by the post-Moi Kibaki administration, estimated the loss at approximately KSh 158.3 billion in 1990s currency — on some calculations close to 10 percent of GDP for the relevant years [TBD-VERIFY: estimates range from roughly KSh 50 billion to over KSh 600 billion depending on interest and currency adjustments]. The Commission found Pattni the principal beneficiary, identified then-Vice-President George Saitoti as having provided Treasury-level facilitation, and named Cabinet Minister Nicholas Biwott as a political protector; it concluded that the scheme could not have run as long as it did without senior political protection (KE-A-02). For the reform movement, Goldenberg was the supreme exhibit: proof that the one-party patronage state, under donor-imposed austerity, had captured the central bank's foreign-exchange machinery to enrich its principals while ordinary Kenyans bore the cost of structural adjustment.

That economic context — stagnant or declining real incomes, rising youth unemployment, and visible elite enrichment — supplied the grievance fuel for renewed mobilisation. From 1995 the constitutional-reform demand re-intensified under a coalition of churches (the NCCK, the Kenya Episcopal Conference, the Supreme Council of Kenya Muslims), legal-profession bodies (the LSK, the International Commission of Jurists Kenya Section), and a constellation of new civil-society organisations: the Citizens Coalition for Constitutional Change (the "4Cs"), the National Convention Assembly (NCA), and its executive arm the National Convention Executive Council (NCEC), the latter chaired by figures including Kivutha Kibwana and drawing on Willy Mutunga and a wide activist network. The NCEC's strategy combined mass action with a programmatic demand for "minimum reforms" before the 1997 election and comprehensive constitutional review thereafter. The mobilisation produced the "Saba Saba II" protests of July–August 1997 — explicitly invoking the 1990 rally — which combined demands for constitutional reform with concerns about the conduct of the impending poll, and which the state again met with force, including the notorious assault by police on demonstrators sheltering in Nairobi's All Saints Cathedral on 7 July 1997 [TBD-VERIFY: the casualty figures for the July 1997 protests are reported variably]. It was this pressure that forced the limited reforms of late 1997, to which the next section turns.

---

11. The IPPG Reforms and the 1997 Election

The Inter-Parties Parliamentary Group (IPPG) reforms of August–November 1997 were the Moi-era regime's negotiated response to the NCEC-led mobilisation. Rather than concede the comprehensive constitutional review the NCEC demanded — and which would have proceeded outside parliamentary control — the government drew the parliamentary opposition parties into a bargained "minimum reform" package agreed within Parliament. The IPPG package included: a partial reconstitution of the Electoral Commission of Kenya to admit opposition-nominated commissioners; repeal or amendment of selected colonial-era public-order statutes, including provisions of the Chiefs' Authority Act and restrictions on public meetings; incremental media liberalisation; and registration of additional parties for the impending poll. The reforms were real but limited: the imperial presidency, the executive's residual control over the electoral process, and the underlying constitutional architecture all remained.

The IPPG bargain split the reform movement, and the split was itself significant. The parliamentary opposition accepted the minimum-reform package and contested the December 1997 election under it; the NCEC and much of organised civil society rejected the package as a co-optation that traded comprehensive reform for cosmetic concessions and that drew the opposition into legitimising an unreformed system. This cleavage — between an electoral-political opposition willing to compete within partially reformed rules and a civil-society movement insisting on prior constitutional redesign — would structure the constitutional-review contest of the following years.

The 29 December 1997 election reproduced the 1992 pattern. Five principal opposition candidates split the anti-KANU vote: Mwai Kibaki (DP); Raila Odinga, now leading the National Development Party (NDP) after splitting from FORD-Kenya; Charity Ngilu (Social Democratic Party), the first woman to contest the Kenyan presidency; Michael Wamalwa Kijana (FORD-Kenya); and others. Moi was returned with 2,500,856 votes (40.40 percent), against Kibaki's 1,911,742 (31.49 percent), Odinga's 667,886 (10.92 percent), Wamalwa's 505,704 (8.29 percent), and Ngilu's 488,600 (7.78 percent) [TBD-VERIFY: precise tallies from the Electoral Commission Final Report and Hornsby (2012)]. Moi again cleared the provincial threshold; Kibaki met it only in Central. Pre- and post-election violence recurred, this time concentrated on the Coast around Likoni, where armed "raiders" attacked up-country settlers in August 1997; the Akiwumi Report subsequently documented the violence and the state's inadequate response [TBD-VERIFY: 1997–98 Coast casualty figures range from roughly 100 to over 200], and again no senior figure was prosecuted. The 1997 result confirmed for the reform movement that no amount of electoral reform short of constitutional redesign would dislodge KANU while the opposition remained divided — a conclusion that pushed the katiba question to the centre of the post-1997 struggle.

---

12. The Constitutional-Review Contest — Ufungamano and the CKRC

The IPPG package had included a commitment to constitutional review, enacted as the Constitution of Kenya Review Act 1997 (amended 1998). But the question of who would control the review — Parliament and the executive, or the civil-society movement that had forced the issue — produced a rivalry that defined the 1998–2001 period and shaped the eventual 2010 Constitution.

The government-sanctioned track was the Constitution of Kenya Review Commission (CKRC), to be established under the Review Act. Civil society, the churches, and much of the opposition rejected the initial framework as government-controlled and insufficiently participatory, and in 1999 convened a parallel process — the Ufungamano Initiative, named for the Ufungamano House venue in Nairobi where it was based and led by religious leaders including the heads of the main Christian councils and the Supreme Council of Kenya Muslims, together with NCEC-aligned activists. Ufungamano established its own People's Commission of Kenya under the chairmanship of Oki Ooko Ombaka [TBD-VERIFY: the precise leadership composition of the Ufungamano People's Commission], advancing a "people-driven" model of constitution-making in opposition to the "parliament-driven" model the government preferred.

The deadlock was broken by merger. After negotiation, the two tracks combined in 2001 into a single, expanded Constitution of Kenya Review Commission chaired by the constitutional scholar Yash Pal Ghai, who brought comparative-constitutional experience including work on the Hong Kong Basic Law. The Ghai-led CKRC undertook extensive public consultation and produced a draft — the "Ghai Draft," refined at the Bomas of Kenya national constitutional conference convened from 2002 — that proposed a substantial dispersal of presidential power, a prime-ministerial office, and devolution. The reform movement's foundational documents on this process are Yash Pal Ghai and Jill Cottrell Ghai's Kenya's Constitution: An Instrument for Change (2011) and the wider Katiba Institute corpus.

The crucial point for the Second Liberation is that the katiba project was not completed in its own decade. The Bomas process continued past the 2002 alternation; the resulting draft was contested and revised by the post-2002 government into the "Wako Draft," which voters rejected at the November 2005 referendum (KE-B-02, when written); and only after the trauma of the 2007–08 post-election violence did a reconstituted process produce the 2010 Constitution (KE-A-04), which finally entrenched the dispersal of presidential power, the independent commissions, the Bill of Rights, and the 47-county devolution that the 1990s reformers had demanded. The Second Liberation thus achieved multipartism in 1991 and alternation in 2002 but bequeathed its constitutional ambition to the following decade.

---

13. The 2002 Alternation — NARC and the End of KANU's 39-Year Rule

The Second Liberation's culminating electoral event was produced by Moi's own succession miscalculation. Barred by the two-term limit he had accepted at the 1991 repeal from standing again, Moi imposed Uhuru Kenyatta — son of the founding President, KANU youth chairman, and a relatively junior politician — as the KANU presidential candidate in October 2002, passing over senior aspirants including Vice-President George Saitoti, Musalia Mudavadi, Kalonzo Musyoka, and Raila Odinga (who had merged his NDP into KANU in March 2002 on the implicit expectation of senior succession consideration). The imposition's competing rationales — Moi's wish to preserve post-retirement influence and shield his network from accountability; KANU's internal factional contest; and the calculation that a Kikuyu candidate could split the Kikuyu vote — are analysed in KE-A-02.

The imposition shattered KANU. Raila Odinga led a "Rainbow Alliance" of disaffected senior KANU figures — Saitoti, Mudavadi, Musyoka, Moody Awori, Joseph Kamotho, and others — out of the party. The Rainbow Alliance, organised as the Liberal Democratic Party, fused with Mwai Kibaki's National Alliance Party of Kenya (itself a coalition of DP, FORD-Kenya, and Ngilu's constituency) to form the National Rainbow Coalition (NARC) in October 2002. An October 2002 Memorandum of Understanding committed Kibaki to create a prime-ministerial post for Raila and to share Cabinet positions among the constituent parties (the contested post-2002 implementation is documented in KE-B-01). At a mass rally at Uhuru Park, Raila Odinga delivered the coalition's defining moment, declaring "Kibaki Tosha" — "Kibaki is enough" — and committing the Rainbow constituency to Kibaki's candidacy.

NARC achieved precisely the condition that the strategic-calculation reading of 1991 had identified as the only one under which KANU could lose: a consolidated, multi-ethnic opposition. In the 27 December 2002 election — administered by the IPPG-reformed Electoral Commission under Samuel Kivuitu and rated credible by domestic and international observers — Kibaki defeated Uhuru Kenyatta 4,501,428 votes (62.2 percent) to 1,835,890 (31.3 percent), with Simeon Nyachae a distant third (KE-A-02, KE-B-01). Moi conceded promptly, and on 30 December 2002 he handed power to a wheelchair-bound Kibaki — recovering from a road accident — at Uhuru Park, with Raila Odinga among the senior NARC figures present. KANU's 39-year continuous dominance, dating from the 1963 independence settlement, had ended through the ballot box rather than through coup, civil war, or constitutional rupture.

The composition of the 2002 victory is itself analytically significant. NARC assembled the broadest multi-ethnic coalition in Kenyan history since the immediate post-independence period: Kibaki's Kikuyu base, Raila Odinga's Luo constituency, Charity Ngilu's and Kalonzo Musyoka's Kamba support, Musalia Mudavadi's and Michael Wamalwa's Luhya following, and a substantial cross-ethnic urban and youth vote exhausted by twenty-four years of KANU rule and a decade of economic stagnation. The coalition succeeded by holding together, for one election, precisely the constituencies whose fragmentation across FORD-Asili, FORD-Kenya, the DP, the NDP, and the SDP had delivered Moi his pluralities in 1992 and 1997. The structural lesson the Second Liberation had learned through two lost elections — that only a consolidated opposition could win under the inherited rules — was applied with discipline in 2002, and the discipline was supplied largely by Moi's own succession blunder, which drove the defecting KANU heavyweights into Kibaki's camp and gave the coalition both its breadth and its momentum.

It is also worth recording what 2002 did not change, because the omissions frame the contested-meaning debate. The election was conducted under the same constitution the reformers had spent a decade seeking to replace; the imperial presidency Kibaki inherited was the one Moi had wielded; the Provincial Administration inherited from the colonial state (KE-A-01) remained intact; and the patronage logic that bound ethnic-regional coalitions to state resources was untouched. The NARC Memorandum of Understanding — the instrument meant to discipline the new coalition's internal power-sharing — had no constitutional standing and collapsed within months (KE-B-01), reproducing in the reform camp the very winner-take-all dynamic the movement had condemned in KANU. These continuities are why the meaning of 2002 is contested.

The meaning of 2002 is itself contested, and the contest is the document's central forward-looking question. On the consummation reading, the activist Second Liberation — the detentions, Saba Saba, the church witness, the LSK litigation, FORD, the NCEC mobilisation — finally bore its fruit in the first peaceful inter-party transfer of power in Kenyan history; the alternation vindicated the decade of struggle. On the elite-alternation reading, 2002 was a recomposition of the political class that left the underlying patronage-state, the imperial presidency, and the ethnic-coalition logic of Kenyan politics fundamentally intact — a reading reinforced by the rapid collapse of the NARC MoU, the unprosecuted Anglo Leasing scandal of the Kibaki years, and the 2007–08 violence (KE-B-01, KE-B-03). The corpus holds that both readings capture valid dimensions: 2002 was simultaneously the Second Liberation's democratic triumph and a demonstration that alternation without constitutional redesign reproduces the inherited architecture — which is why the katiba project remained the unfinished business carried into the 2010 Constitution (KE-A-04).

---

14. Conclusion — The Second Liberation as Unfinished Project

The Second Liberation occupies a distinctive place in the Kenyan governance archive: it is the period in which the post-colonial one-party state was forced open, but not the period in which a new constitutional order was built. Three conclusions structure its legacy.

First, the era established the constitution as the master category of Kenyan reform politics. Because the one-party state had foreclosed lawful opposition through constitutional amendment, the reform movement framed its struggle as a katiba struggle from the outset, and that framing outlived the 1991 repeal and the 2002 alternation. The independent commissions, the entrenched Bill of Rights, the dispersed presidency, and the devolved counties of the 2010 Constitution (KE-A-04) are the delayed achievements of a movement whose demands were articulated in the 1990s but whose institutional realisation required the trauma of 2007–08 and a further three years of negotiation.

Second, the era demonstrated the limits of procedural democratisation under inherited rules. The reformers won multipartism in 1991 and discovered in 1992 and 1997 that multipartism alone, under first-past-the-post with provincial thresholds and against a fragmented opposition, reproduced incumbent rule. The lesson — that the rules of the game, not merely the right to play, must be redesigned — is the era's principal contribution to the comparative study of democratic transitions, and it explains why the Kenyan reform movement was, uniquely among African pro-democracy struggles of the period, so persistently focused on constitution-making rather than on elections alone.

Third, the era's failures of accountability cast a long shadow. The unprosecuted Rift Valley and Coast clashes, the unresolved Ouko murder, the Nyayo House torture practices, and the Goldenberg scandal were all documented — by Divide and Rule, by The Cursed Arrow, by the Akiwumi and Bosire Commissions, and by the post-2003 Truth Justice and Reconciliation Commission — but none produced senior-level legal consequence. This pattern of documentation-without-accountability is widely read as the antecedent of the 2007–08 post-election violence (KE-B-03), in which ethnic violence again served electoral strategy, and as a structural feature of Kenyan governance that the 2010 Constitution sought, with mixed success, to break.

The two contested-record positions on the Moi era (KE-A-02) bear directly on the Second Liberation. The institutional-stability reading credits Moi with managing the transition without state collapse and ultimately conceding power peacefully; the institutional-damage reading treats the patronage-state, the ethnic-violence instrumentalisation, and the Goldenberg-scale corruption as the era's defining legacy. The Second Liberation narrative sits between them: it records both the genuine heroism of those who forced the system open and the structural continuities that survived the opening. Whether the activist heroism, the donor pressure, or the elite calculation deserves primary credit for the liberalisation; whether the 1990s clashes were spontaneous or orchestrated; and whether 2002 was consummation or mere alternation — these remain open questions in the historiography, and the corpus preserves them as open rather than resolving them.

---