Round 2 Critic — Framing A (Institutional Capture / ANC Patronage)
Steelman: Framing B (Procedural Legitimacy) vs Framing A
The strongest version of Framing B runs as follows: the 2009–2018 protection gap, when examined in granular legal terms, is substantially explicable without invoking corruption as a cause. Mpshe’s April 2009 decision to drop charges was not a backroom favour — it was a formal prosecutorial discretion decision based on a legal assessment that the Spy Tapes demonstrated the timing of Zuma’s charges had been politically coordinated. That assessment may have been wrong; it was litigated aggressively; but the Constitutional Court’s later ruling that the SCA lacked jurisdiction to set it aside is not a minor procedural technicality. It means the highest court in the country effectively held that Mpshe’s decision, however flawed, fell within the lawful exercise of prosecutorial discretion. Framing A has to grapple with this: if the ConCourt said the decision had standing, calling it straightforward “capture” requires an additional step — showing that Fraser’s provision of the tapes was itself an act of corruption that Mpshe knew or should have identified. The vault does not contain evidence that Mpshe was personally corrupt rather than wrong. Framing B can claim that distinction matters.
The 2018–2026 delay is even more legible under Framing B. Once charges were reinstated in 2018, Zuma’s legal team pursued every available constitutional avenue: s 35 fair trial challenges, application to stay proceedings on grounds of prosecutorial delay, health-based adjournments, the Stalingrad strategy of review applications against individual interlocutory decisions. None of this is aberrant. It is the ordinary exercise of rights guaranteed by Chapter 2 of the Constitution. The arms deal case involves hundreds of thousands of pages of documentary evidence, international witnesses across five jurisdictions, a French co-accused (Thales), and allegations spanning three decades. A six-year trial preparation and hearing period is not, on its face, scandalous. For context: the Mozambican Tuna Bond case — far simpler in evidentiary terms — took five years to result in convictions in the UK. Framing B’s best argument is that Zuma’s lawyers were effective, constitutional rights are robust, and complexity is real.
The irreducible minimum. After conceding the above, what residue of “scandal” cannot be explained by Framing B?
First: the NPA’s use of prosecutorial energy against Pravin Gordhan (under Abrahams, 2017) — charges widely acknowledged as legally hollow — while Zuma’s far more substantiated arms deal charges sat dormant. Framing B has no account of this. Prosecutorial discretion cannot explain a decision to charge the Finance Minister with fraud based on an intelligence report while simultaneously declining to bring to trial a 783-count indictment with a convicted co-accused (Schabir Shaik). That asymmetry requires a purpose. Framing A provides one; Framing B does not.
Second: Nxasana. His paid exit — R17.3 million, later ruled unlawful — is not a procedural delay. It is the removal of a functionary who was, by all accounts, preparing to prosecute Zuma. You cannot square Nxasana’s departure with Framing B unless you accept that a government legitimately purchased the resignation of its chief prosecutor to avoid a constitutional conflict. If you accept that framing, you have essentially conceded the capture narrative in politer language.
Third: Menzi Simelane’s appointment, invalidated by the Constitutional Court for misrepresentation during his vetting process. The court found he lied. This is not procedural delay. It is a captured appointment that the appointing authority — Zuma — had to have known was compromised. Framing B cannot absorb this without becoming incoherent.
The minimum residue sufficient to sustain Framing A: two unlawful or invalidated NDPP appointments (Simelane, Abrahams), one paid unlawful removal (Nxasana), and one documented asymmetric use of prosecutorial power (Gordhan charged / Zuma dormant). None of this is explained by constitutional-design complexity. All of it is explained by capture. Framing B survives as a partial account of the delay; it fails as a complete one.
Political Convenience Hypothesis at Full Strength
Do not mistake the current Zuma trial for the restoration of accountability. Make the strongest possible case for the political convenience hypothesis and it runs as follows.
Charges were reinstated against Zuma in March 2018 — six weeks after Cyril Ramaphosa was elected ANC president at Nasrec in December 2017. The timing is not proof of bad faith, but it is striking enough to warrant explanation. The NPA’s position from 2009 to 2017 was that the charges were either dormant (under Simelane, Nxasana) or being actively redirected (Abrahams). The moment Zuma’s internal ANC opponent assumed the ANC presidency, the charges were reinstated. If you are committed to the Framing B story — the NPA is an independent institution making legal decisions — you need to explain why the legal landscape changed so dramatically in six weeks when the evidentiary record had not.
By 2026, the political stakes are sharper still. Zuma now leads the uMkhonto weSizwe (MK) party, which in the May 2024 general election took 14.6% of the national vote, cost the ANC its outright parliamentary majority, and is particularly dominant in KwaZulu-Natal, Zuma’s home province. The ANC governs only through the Government of National Unity because MK split its traditional base. Zuma’s conviction — or even his continued subjection to a criminal trial — neutralises the ANC’s most effective internal rival without requiring a single vote. If the NPA were independently pursuing the Zuma prosecution purely on legal grounds, you would expect to see equivalent energy directed at figures whose prosecution costs the current coalition. You do not see that. Mantashe, Zondo Vol 3 referral for Bosasa: no charges. Mokonyane, Zondo Vol 3 referral: no charges, subsequently elected ANC DSG. Fraser, Zondo referral on multiple counts: prosecution declined June 2024.
The “accountability restored” narrative under Ramaphosa is, on this reading, structurally identical to the “captured NPA” of the Zuma era — just with different beneficiaries. The Zuma-era NPA captured energy toward protecting Zuma and attacking Gordhan. The Ramaphosa-era NPA appears to direct energy toward prosecuting Zuma, Magashule (Zuma faction), and Matodzi (state capture but outside current coalition) while declining Phala Phala, not pursuing CR17, not charging Mantashe, and declining Fraser. The pattern in both eras is not “the NPA prosecutes when there is evidence”; it is “the NPA prosecutes who the current faction needs prosecuted.”
The reform institutions — Batohi as NDPP, IDAC made permanent in August 2024, FATF compliance — do not address this structural problem. They address capacity and resource constraints, not the appointment architecture. NDPP Batohi was appointed by Ramaphosa. Batohi’s successor Andy Mothibi (February 2026) was appointed by Ramaphosa’s President-of-the-Republic successor within the same GNU. The NPA Amendment Act 2024 that made IDAC permanent did not change how the NDPP is appointed: still by the President, on advice of the JSC. A president who needs the NPA to proceed on certain targets and decline on others retains, through the appointment power, the structural lever to achieve exactly that outcome across multiple NDPP cycles — without ever issuing a single direct instruction to any prosecutor.
FATF compliance is surface legitimation. South Africa exited the FATF grey list in October 2024 based on demonstrated progress across 22 action items, most of which measure process outputs (asset restraint orders, case registrations, policy frameworks) rather than conviction rates for senior politicians. R10.6 billion in assets under restraint sounds significant; the NPA’s own data shows the bulk of asset restraint involves lower-tier state capture beneficiaries, not the NDPP-level political targets that the Zondo Commission prioritised. You can pass FATF and still have a factionally directed prosecution service. The grey list exit is not evidence of independence.
The Fraser Immunity Test
Arthur Fraser is the single hardest test case for NPA independence under Batohi, and it has no satisfactory answer within the “accountability restored” narrative.
Fraser’s exposure to prosecution is not based on political accusation. It rests on: (a) the Zondo Commission’s formal referral for criminal investigation on at least three distinct grounds — SSA misuse including Project Wave and cash payments to Zuma, the Principal Agent Network’s misappropriation of approximately R1.5 billion in public funds, and the unlawful medical parole of Zuma; (b) the Supreme Court of Appeal’s ruling that his parole decision was unconstitutional and unlawful — a judicial finding, not an allegation; (c) criminal charges for defeating the ends of justice filed by Paul O’Sullivan arising from the parole decision; (d) DA criminal charges relating to the Spy Tapes, including the allegation that some NIA recordings were obtained and leaked in breach of intelligence secrecy obligations.
The NPA’s response to this body of evidence, across multiple overlapping complaints spanning the full Batohi era: declined prosecution in June 2024 on the O’Sullivan charges (Phala Phala / parole nexus); no prosecution decision on Zondo referral matters as of April 2026, with NDPP Batohi having retired January 30, 2026 without having filed charges; new NDPP Mothibi has given no indication of reopening Fraser files.
The significance is structural, not merely factual. Fraser is not a peripheral figure. He is, by the vault’s own account, the actor most directly responsible for the 9-year protection gap: his Spy Tapes manipulation in 2009 provided the legal pretext for Mpshe’s charge withdrawal, and that withdrawal — whatever its formal legal standing — is the event from which the entire gap flows. If the reformed NPA is genuinely independent and accountability-oriented, prosecuting Fraser is not a close call. It is the clearest case in the post-Zondo referral set. The Zondo findings on Fraser are more specific, more substantiated, and more directly linked to judicial findings of unlawfulness than the equivalent findings on Mokonyane or Mantashe.
Yet the NPA declined. No explanation meeting evidentiary standards has been published. The contrast with the NPA’s willingness to proceed on Zuma (whose charges involve older, partly overlapping evidentiary terrain) is not explicable on legal grounds alone.
The Fraser immunity problem is left standing as an unanswered objection: if the reformed NPA declines to prosecute the man the Zondo Commission said engineered the original protection gap, on what grounds does the “accountability restored” claim rest?
What Framing A Survives After This Round
Framing A (Institutional Capture / ANC Patronage) survives this round in a more precise and harder-edged form than Round 1 left it.
Its strongest surviving claim is not that the NPA was captured under Zuma — that is settled. It is that the NPA’s selectivity under Ramaphosa follows the same structural logic as capture, even if its personnel are more legally competent and its processes more formally independent. The capture under Zuma was crude: invalid appointments, paid removals, explicit redirections. The selectivity under Ramaphosa is structural: correct appointments made by an executive with political interests, operating within an appointment architecture that does not require explicit direction to produce convenient outcomes.
The VBS/Matodzi comparison survives. The argument that “systemic impossibility” explains the Zondo implementation gap is falsified by Matodzi’s conviction in under three years under the same pipeline. Framing B cannot use complexity as a blanket defence when the NPA demonstrates it can move quickly when the target is outside the current coalition.
The Gordhan-charged / Zuma-dormant asymmetry survives as the clearest single piece of evidence that the Zuma-era NPA was not merely slow or under-resourced but actively redirected. No Framing B account explains it without conceding the point.
The reform-ceiling argument survives in its sharpest form: Ramaphosa as ANC Deployment Committee chair 2012–2017 during the appointment of Shaun Abrahams, the entrenchment of Nomgcobo Jiba, and the Moyane-SARS capture is not a peripheral biographical detail. It means the current reformer was a participant in the decisions that created the problem he is now credited with fixing. The reform is real at the margins and insufficient at the structure.
What Framing A cannot claim after this round: that Mpshe was personally corrupt (unproven), that the entire 2018–2026 delay is manufactured (significant portions are attributable to legitimate legal process), or that constitutional complexity is never a real factor.
Two Final Unanswered Objections
Objection 1: The falsifiability problem of factional prosecution. The political convenience hypothesis — that the Zuma prosecution is factionally motivated rather than accountability-driven — cannot be falsified from vault data. If true, the distinction between “capture” and “factional prosecution” collapses the very accountability framing the post-Zondo narrative depends on. But if the hypothesis is false, and the NPA is genuinely independent, the Fraser non-prosecution and the Ramaphosa-touching non-actions require an evidentiary explanation the NPA has not published. The vault does not contain information that resolves this. The recon session leaves open whether South Africa is experiencing genuine accountability restoration or factional accountability — a prosecution regime that targets enemies of the current coalition while protecting its members — in a form structurally indistinguishable from what it replaced.
Objection 2: No constitutional path to structural reform. The NDPP appointment architecture is the root cause identified by Round 1 and confirmed by the analysis above. No credible political path to reforming it exists within the current GNU. Changing the appointment mechanism requires constitutional amendment (two-thirds majority in the National Assembly) or, at minimum, a restructured NPA Act requiring JSC endorsement to carry meaningful veto weight. The ANC, even in coalition, does not have a two-thirds majority. The DA, the largest GNU partner, has not campaigned on NDPP appointment reform. The MK party and EFF, who together might complete a blocking minority on other issues, have no incentive to strengthen the prosecution service that is currently pursuing Zuma. No civil society proposal for structural reform of the appointment architecture has received parliamentary traction. The Zondo Commission’s recommendation on this point — that appointment should require broader parliamentary endorsement — has not been legislated. The reform ceiling is not a temporary political constraint; it is baked into the current parliamentary arithmetic for at least the duration of the 2024–2029 term. Nothing in the vault suggests a mechanism by which this changes.
Timing: Started 2026-04-17 12:01:08 · Finished 2026-04-17 12:08:41