In September 2025, the High Court reinstated corruption charges against a sitting Cabinet Secretary after finding that the Office of the Director of Public Prosecutions had withdrawn them without consulting the Ethics and Anti-Corruption Commission (EACC) and outside its own internal guidelines.^[1] The EACC has told the Senate that eighteen high-profile corruption cases have been withdrawn, although accounts differ on when that count begins.^[2] This raises an obvious question. Before a withdrawal takes effect, a court must agree to it. What should that agreement require?

Kenya’s Constitution requires the court’s permission before the Director of Public Prosecutions (DPP) discontinues a prosecution, yet no settled standard tells a magistrate what that permission should involve. This article argues that consent under section 87(a) of the Criminal Procedure Code is the one point at which Article 157(11) can be enforced before a withdrawal takes effect, that the missing standard is a real gap, and that needs to be addressed.

The question is not only for lawyers or legal professionals. Corruption cases concern public money, and when charges are dropped the public may never see a trial or verdict. The court’s consent is the last public checkpoint before that happens. The same power also touches ordinary accused persons who can be discharged without acquittal and charged again years later. How courts use that checkpoint therefore shapes whether Kenyans can trust that cases are dropped for good reasons and not for convenient/biased ones.

The Legal Framework

Section 87(a) of the Criminal Procedure Code affirms that Kenya’s courts must consent to prosecutorial withdrawals but lack any settled standard for doing so, therefore weakening Article 157(11) of the constitution.

Section 87(a) of the Criminal Procedure Code lets a prosecutor withdraw a charge in a subordinate court with the court’s consent. If this happens before the accused is called upon to make a defence, the result is a discharge, not an acquittal, and the charge may be revived.^[3] The Code was enacted in 1930, under British colonial rule, and although it has been amended over the years it has never been rewritten for the constitutional order that came with independence, or the one adopted in 2010. Commentators have argued that the law of prosecution built around it needs fundamental rethinking. If that is right, the difficulty is not confined to the laws derived from the Code. Section 87(a) sits inside an instrument that may itself need rewriting, not merely patching.^[4]

The 2010 Constitution changed the setting, article 157(10) shields the DPP from direction or control, while Article 157(11) requires regard to the public interest, the interests of the administration of justice, and the need to avoid abuse of legal process. Two further clauses speak directly to withdrawal: Article 157(8) forbids discontinuance without the court’s permission, and Article 157(7) requires acquittal if proceedings are discontinued after the close of the prosecution’s case.^[5] That second clause sits uneasily with section 87(a), which allows mere discharge at any point before the defence is called. The ODPP Act operationalises the office’s powers,^[6] and the office has published guidelines on the decision to charge.^[7] There is, however, no equivalent guidance on the decision to withdraw.

Two Judicial Signals

Two authorities pull in different directions. In Republic v Muneh Wanjiku Ikigu, the High Court declined to allow a withdrawal that would have left the accused exposed to indefinite re-prosecution, observing that the power to withdraw or initiate cases can no longer be used “whimsically and arbitrarily.”^[8] In Diamond Hasham Lalji, the Court of Appeal held that prosecutorial discretion is reviewable but not absolute, that courts should not unjustifiably interfere with it, and that the burden shifts to the DPP only once sufficient evidence of a breach is shown.^[9] Some High Court revision rulings have since begun to describe what consent requires, but they come from individual judges and, so far as can be told, have not been consolidated into a settled test.

Consent in Practice

The available evidence is anecdotal but suggestive. In February 2024, a magistrate refused to allow withdrawal of a Ksh 30 million case against a former Kenya Pipeline Corporation managing director.^[10] In April 2026, a Kisumu magistrate allowed a section 87(a) withdrawal after noting Article 157(8) and the conditions the accused had agreed to.^[11] Both outcomes may be defensible. What is missing is a visible, shared logic that explains why one application fails and another succeeds. No systematic, court-record-based study yet tests how consent is handled across stations, and the figures in circulation come largely from advocacy and press sources.

For ordinary Kenyans these inconsistencies are not abstract. A withdrawn corruption case means allegations about public money that may never be tested in open court, while a discharged accused person may live for years under the shadow of a revived charge. In both situations, the reasons a court asks for before consenting decide whether the system looks fair or arbitrary.

The Constitutional Lens

Okoth-Ogendo’s thesis of “constitutions without constitutionalism” describes constitutional texts that promise limits on power while practice leaves that power largely unconstrained.^[12] Gayoye’s application of the thesis to Kenya’s judiciary suggests that whether text becomes practice depends heavily on the willingness of individual judges to give it effect.^[13] Article 157(11) is a textbook candidate for this paradox. Other writers have looked for checks elsewhere, for instance by exploring private prosecution as a corrective to decisions not to prosecute,^[14] and officials within the ODPP have defended the office’s mandate to review and withdraw matters under the same constitutional considerations.^[15]Put together, the position is this: the Constitution makes the court’s permission a requirement, and the statute gives the court a consent function,^[16] but neither says what the court should examine before it consents. The standard to guide the procession from required duty to the granting or refusal of consent for withdrawal of a case is missing. The ODPP’s published guidance addresses charging, not withdrawal. The two leading precedents set a tension without resolving it: courts must not be passive, as the Ikigu case insists, yet they must not unjustifiably interfere, as the Lalji case cautions. The result is that consent may operate as a real constitutional check in one courtroom and as a formality in another, and neither the ODPP, the EACC, complainants nor accused persons can predict which. The mismatch between Articles 157(7) and 157(8) and the discharge mechanism in section 87(a) compounds the uncertainty.

A Consent Structure Worth Adapting

What should fill the gap? A workable structure ought to be elaborate. The five steps below are a starting point, drawn from the constitutional text. They are offered to be adapted by judges, the ODPP or Parliament, not as a finished code.

1. Reasons on the record. The prosecutor states why withdrawal is sought and ties those reasons to the considerations in Article 157(11): the public interest, the interests of the administration of justice, and the avoidance of abuse of legal process. A bare assertion that withdrawal is in the public interest should not be enough.

2. A basis in the material. The court asks whether the stated reasons are supported by what is before it, such as the state of the evidence or the ODPP’s own guidelines. 

3. Consultation and a chance to be heard. Where an investigating agency or complainant is affected, the prosecutor shows they were consulted, and the court hears them if they object. 

4. Consequences and timing. The court considers what withdrawal means for the accused, including the risk of indefinite re-prosecution and how the stage of the case interacts with the acquittal rule in Article 157(7).

5. A reasoned ruling. The court records briefly why it consents or refuses, so that the decision can be reviewed and the public can understand it.

The structure can be scaled. For routine, early-stage applications a short statement and a one-paragraph ruling may suffice. For late-stage cases, prosecutions involving public money, or contested applications, courts would expect fuller reasons and a hearing for affected parties. The structure could be adopted through judicial practice directions, through ODPP withdrawal guidelines modelled on its charging guidelines, through a targeted amendment to section 87 or, more ambitiously, through a rewrite of the Code itself, and it should be tested against how courts actually behave today.

Conclusion

Circling back to the Cabinet Secretary case with which this article began. The ODPP withdrew the charges without consulting the EACC and outside its own guidelines, and it took the High Court to restore them. If eighteen high-profile corruption cases have indeed been withdrawn, each one raises the same question: did the court that consent ask the right questions? Without a settled standard, nobody, whether a citizen, a complainant or an accused person, can say.

A clear standard, however if developed, would matter beyond the courtroom. For citizens, it would mean that dropping a corruption case has to be explained and tested before it takes effect, not discovered afterwards. For accused persons, it would mean that discharge is not left hanging over them without reasons. For the ODPP and the EACC, it would mean predictable rules in place of repeated public disputes. The absence of such a standard leaves the constitutional promise of accountable prosecutorial discretion dependent on the habits of individual magistrates. 

Allbright Mokeira Kaggia is a researcher and LLB finalist at Kabarak University School of Law, with a growing interest in governance and human rights.

Endnotes