Reserve Bank of India (Commercial Banks – Governance) Amendment Directions, 2026 dated 14-July-2026

The single most consequential structural change is the arrival of Appendix II B. Until now, a large volume of matters travelled through a committee purely so that they could be tabled at the Board, with the committee adding a recommendation but rarely a decision. With delegation now expressly permitted at the Board’s discretion, committees such as the Audit Committee of the Board (ACB), the Risk Management Committee (RMC), the Nomination and Remuneration Committee (NRC), the IT Strategy and the IT / Cyber Security (CSITE) committees, and the customer service and fraud monitoring committees become the terminal decision-making forum for the matters delegated to them, rather than a waypoint on the route to the Board.

That change is more demanding than it first appears. Committee charters drafted in the language of ‘review and recommend to the Board’ will potentially need to be rewritten in the language of decision rights, with explicit financial and qualitative thresholds, named escalation triggers that pull a delegated matter back to the Board, and a defined standard of reporting upward. Committee composition and skills also come under pressure: a committee that decides needs sufficient independent membership, genuine subject-matter depth (particularly on technology, cyber, model risk and climate), and a chair willing to be accountable for what the Board did not see.

We expect three second-order effects. First, the centre of gravity of governance debate shifts to the committee room, which raises the importance of committee minutes as the primary evidence of oversight. Second, cross-cutting risks will outgrow single-committee mandates, and leading Boards will convene joint sessions, for example the ACB with RMC on credit and provisioning judgements, RMC with the IT / CSITE committee on cyber resilience and AI model risk. Third, committee load will rise materially, and Boards will need to guard against simply relocating the old checklist one level down.

A principles-based framework only produces better governance if Board behaviour changes with it. The amendment gives the chairperson primary responsibility for the agenda; in practice this means the agenda must be actively curated against the bank’s strategy and risk priorities, rather than assembled by the secretariat from a standing template and last quarter’s carry-forwards.

Several behavioral changes follow. Boards will need to manage time as a scarce governance resource and be able to show how much of it went to strategy and risk governance, satisfying the regulatory floor in paragraph 19B in substance and not merely in the minutes. Meetings will need to move from consumption of management presentations to interrogation of pre-read material, with management expected to bring options, dissenting views and the downside case rather than a paper seeking ratification. Independent Directors, in particular, will need to ask for what is not on the agenda, and to become comfortable exercising judgement on materiality now that no prescriptive list underwrites their decision.

Equally important are the less visible habits: a willingness to defer a decision for a deeper look rather than clear the agenda; regular executive sessions without management present; a culture in which constructive dissent is recorded rather than smoothed over; and an acceptance that under a principles-based regime, the documented rationale for a judgement becomes the compliance artefact. Board evaluation criteria and the NRC’s skills matrix will need to be recast accordingly, rewarding quality of challenge rather than attendance and coverage.

In our experience with comparable governance transitions, the likeliest failure mode is a re-labelling exercise: the secretarial team maps the existing agenda onto Appendices I, II A and II B, the charters are updated, and nothing about how the Board spends its time actually changes. A second and related failure mode is defensive conservatism i.e. retaining everything at the Board because delegation feels supervisory riskier than over-inclusion, which forfeits the entire benefit of the amendment.

Banks should also plan for the following practical frictions:

• Evidencing judgement: with the prescriptive list withdrawn, supervisory engagement will test the reasoning behind what was reserved and what was delegated. Banks need a defensible, dated and minuted basis for those calls, and a periodic review trail, rather than a one-time mapping document.
• Overlapping regimes: the Board’s reserved-matters framework must sit consistently alongside the Companies Act, 2013 (including sections 179 and 180 and the Secretarial Standards), SEBI LODR obligations for listed banks, and RBI’s other circular-level requirements, several of which were written on the assumption of mandatory Board tabling.
• Entity-specific constraints: PSBs operate within statutory schemes and with government-nominee directors that limit delegation flexibility; foreign bank branches must reconcile the framework with local management committee structures and head-office reporting lines; SFBs, PBs and LABs face the same architecture with materially thinner Board and committee bandwidth.
• MIS built for the wrong purpose: existing management information is designed to populate Board packs, not to report back on delegated matters by exception. Without redesign, delegation creates an oversight blind spot rather than headroom.
• Independent director capacity: a small pool of independent directors now carries decision-making committees rather than recommending ones, with consequent demands on time, induction and continuing education.
• A short runway: with effect from 1 October 2026, most banks have only one or two Board cycles in which to approve revised charters, the delegation-of-authority matrix and the agenda calendar, alongside the usual quarterly workload.

A delegation-based framework generates governance data i.e. what was reserved, what was delegated, what came back, how long it took and how Board time was spent and that data cannot be managed on spreadsheets and email. We see four areas where tooling materially improves the quality of Board oversight under the amended directions:

• Board portal as the system of record: a secure portal holding the reserved-matters and delegation register mapped line-by-line to Appendices I, II A and II B, with version history, approval trails, committee decision logs and evidence of the periodic review contemplated by paragraph 19B.
• Agenda and time analytics: dashboards that measure agenda mix against the strategy and risk governance floor, track time-on-topic, deferrals and repeat items, and give the chairperson objective input into agenda design instead of custom and precedent.
• Regulatory change management: automated ingestion and classification of RBI and other regulatory communications, with rules-based routing to the Board, the relevant committee or management. This replaces the blanket tabling requirement of the deleted paragraph 17 with a triaged, auditable flow and is the control that keeps rationalisation from becoming omission.
• AI-assisted Board reading and enquiry: large language model tools that condense long Board packs into issue briefs with linked source references, surface changes since the previous submission, draft candidate questions for directors, and allow natural-language search across the Board pack and minutes archive (‘what did we conclude on this exposure last year’). Used well, these compress preparation time and raise the standard of challenge.
• Risk and early-warning aggregation: integrated risk data and exception dashboards — credit early-warning signals, model performance, cyber and operational resilience indicators, conduct and customer complaints, reported to the Board and committees on an exception basis, consistent with sound risk data aggregation practice.