Accountability Without Decision Rights Is Organisational Theatre

Many organisations say they want accountability, yet design work so that the person answerable for the result must seek approval from several people who are not answerable for it. The result is predictable: decisions travel upward, meetings multiply, risks are hidden in process and delivery slows while everyone remains busy.

This affects both sectors. A business leader may own a revenue, customer or transformation target but lack authority over pricing, talent or investment. A public manager may be accountable for service delivery while routine operational decisions are delayed by unclear political-administrative boundaries, committee overlap or fear of audit consequences. In both cases, responsibility has been assigned without the operating authority required to fulfil it.

Why decision architecture now matters

South Africa’s Public Service and Administration Budget Vote of 19 May 2026 described reforms intended to separate political and administrative authority, make accountability follow responsibility, reduce political bottlenecks and establish clearer lines of responsibility and consequence management. The principle extends beyond government: clear authority is not the enemy of governance; it is a condition for enforceable accountability. Read the 2026/27 Budget Vote speech.

The OECD’s 2026 Trust Survey found that only 42% of respondents across OECD countries considered government decisions likely to be evidence-informed, while 40% felt that public services adapt to changing societal needs. Positive assessments of evidence-informed decision-making are strongly associated with trust in national government. Decision quality, explainability and responsiveness are therefore not internal management concerns alone; they shape institutional legitimacy. Explore the OECD’s 2026 findings.

The private-sector pressure is equally immediate. In an October 2026 analysis of the “judgment bottleneck,” McKinsey argues that AI can accelerate analysis faster than institutions can exercise judgement and mobilise execution. As information moves faster, unclear decision rights become more—not less—costly. Read the October 2026 analysis.

1. Map critical decisions—not only reporting lines

Organisation charts show who reports to whom. They rarely show who can approve an exception, change a service standard, allocate scarce capacity, accept a risk or stop a failing initiative. Those decisions determine whether strategy moves.

Identify the ten to twenty recurring decisions most consequential to performance. For each, document the trigger, required evidence, decision owner, contributors, deadline and escalation threshold. Start where delay has the greatest cost: customer recovery, procurement, workforce allocation, policy exceptions, project prioritisation and risk acceptance.

2. Name one accountable decision owner

Consultation may be broad, but final accountability should be singular. When several committees or executives can veto a choice while no one owns the outcome, consensus becomes a hiding place. People protect their function rather than optimise the institution.

Name one role—not a meeting—as the decision owner. Define who provides input and who must be informed. The owner must be able to explain the evidence, the trade-offs and the result. Others should challenge constructively without acquiring an informal veto.

3. Delegate through guardrails, not vague empowerment

“Use your initiative” is not a decision system. Leaders hesitate when they do not know the limits of their authority or whether a defensible decision will be supported if the outcome is imperfect. Conversely, delegation without boundaries can expose the institution to uncontrolled risk.

Set explicit guardrails: financial limits, legal and ethical requirements, service standards, stakeholder commitments and risk tolerances. State which conditions require escalation. Within those boundaries, support the authorised leader’s decision. Good governance controls the conditions of choice; it does not centralise every choice.

4. Move routine decisions closest to the work

Senior leaders should focus on choices that shape direction, allocate major resources or create enterprise risk. When executives approve routine operational matters, they create a queue and weaken the judgement of people nearer the customer, citizen or process.

Review approval layers and remove those that add no distinctive evidence, risk protection or statutory value. Give frontline and middle-management roles the data, competence and authority to resolve defined issues. Escalate exceptions, not normal work.

5. Make cross-functional trade-offs explicit

Many slow decisions are not unclear; they are contested. Finance protects cost, operations protects continuity, HR protects people processes, technology protects architecture and risk protects exposure. Each may be rational within its mandate while the enterprise result deteriorates.

Define the shared outcome and the hierarchy of trade-offs before debate begins. Decide which objective is primary, what cannot be compromised and what can flex. Record the rationale so that implementation teams do not reopen the decision when pressure arrives.

6. Measure decision performance

Institutions measure projects, budgets and service levels but often ignore the decision system that produces them. Without evidence, slow governance can look like diligence and repeated reversals can look like responsiveness.

Track decision latency, number of hand-offs, escalation frequency, decisions reopened, time from decision to action and outcomes against the original assumptions. Use a short after-action review for high-impact decisions. The aim is not to punish uncertainty; it is to improve judgement and remove structural delay.

A 30-day decision-rights reset

An executive team can begin without launching another large programme:

1. Select ten recurring decisions that materially affect delivery, risk or stakeholder trust.
2. Track their elapsed time, hand-offs and points of rework for two weeks.
3. Name one accountable owner for each decision and clarify required contributors.
4. Write the authority limits, non-negotiable guardrails and escalation thresholds.
5. Remove one approval layer that contributes neither evidence nor risk protection.
6. Add decision latency and reversals to the next executive performance review.

Accountability is not created by demanding greater ownership in speeches or performance contracts. It is designed through clear authority, usable information, disciplined escalation and visible consequences. When leaders align these elements, the institution can move faster while becoming more—not less—governable.