When we speak about the independence of an anti-corruption commission, we often begin with distance: protection from political direction, fair appointments, secure funding and the freedom to follow evidence without fear or favour. These safeguards are essential, and international standards such as the United Nations Convention against Corruption and the Jakarta Principles are right to insist on them. But what does independence look like when everyone knows everyone?
In small island societies, relationships are not incidental to public life. Family ties, church affiliations, village connections, school networks and professional relationships overlap in ways that are normal and unavoidable. The person making a complaint may be related to the person they are reporting. An investigator may have attended school with a witness. A public official might find that a matter before them involves someone from their community or extended family.
These realities are not unique to the Pacific, but they are more visible in small jurisdictions where social distance is limited. Relationships cannot be eliminated. For anti-corruption agencies, the challenge is to ensure they do not determine public decisions.
This question is not theoretical for me. It comes from my experience helping to establish Tonga’s dedicated anti-corruption institution under the Anti-Corruption Commissioner Act, at a time when Pacific countries are strengthening their wider integrity systems through commitments such as the Teieniwa Vision. What follows is drawn from that institutional practice rather than from interviews or fieldwork, and it makes no representative claim about Tonga or the wider Pacific.
A commission may be legally independent on paper yet still struggle to convince the public that decisions are impartial. An agency may carefully manage conflicts of interest and follow rigorous processes, but if the public cannot see how those safeguards operate, trust may remain fragile. As Luís de Sousa has shown, agencies can hold formal powers and still become institutionally irrelevant. The question is how independence becomes visible and credible.
I describe this as practised institutional independence: the conversion of formal autonomy into decisions that are impartial, explainable and reviewable. The framework I set out here is my own, but three of its four dimensions restate concerns already established in the literature on integrity institutions. Structural authority asks whether an agency has the lawful position, powers, protections and resources to decide independently; these are the benchmarks set out in the Jakarta Principles and elaborated in the UN Office on Drugs and Crime’s Colombo Commentary. Operational capability asks whether it has the people, secure systems and expertise to use that authority, the gap between formal powers and practical effect that de Sousa identifies. Procedural fairness asks whether decisions are consistent, proportionate and supported by recorded reasons, which is where Tom Tyler’s work on procedural justice links fair process to legitimacy and public cooperation. The fourth dimension is the addition I am proposing. Relational integrity asks whether personal, family, community and professional ties have been identified and prevented from controlling those decisions. Weakness in any one can turn formal independence into dependence, arbitrary power or public distrust.
Relationships are not inherently corrupt. Grant Walton and David Jackson’s work on reciprocity networks in Papua New Guinea shows that the same ties can supply social support and enable favouritism. The response is neither to assume that every relationship is evidence of bias nor to ignore relationship-based risk because it is common. OECD guidance on conflicts of interest points the same way: identify, disclose, assess and manage, rather than automatically exclude.
Consider a single file. An officer assigned to a complaint recognises a community connection to an organisation named in it. She discloses the connection and withdraws from the initial assessment. Her access to the file is restricted, and a senior decision-maker records the risk and the control chosen. Depending on the closeness of the tie and the safeguards available, the matter may continue under supervision, be reassigned or be referred for external assistance. The institution has applied a proportionate control and left a record that can be reviewed later, without disclosing the complaint to anyone who does not need to see it.
Recusal should be normal, but it cannot be the only tool. In a small agency, repeated recusals concentrate work in too few hands or stall a matter altogether. Controls therefore need to be graded. An officer with a minor connection might be kept out of the decision while still performing a technical task under supervision. Another may need to leave the matter entirely. Senior leaders should disclose their own connections first, so that staff understand that stepping aside is professional strength rather than embarrassment.
This is also a question of the vā, the relational space between people and institutions that Sione Finau, Melaia Paea and Martyn Reynolds describe as linking relationality, ethics and action. Caring for that space does not mean avoiding hard decisions. It means exercising authority with respect, clarity and consistency. A fair process can preserve a person’s dignity when the outcome is unwelcome, while an unexplained or humiliating one damages trust even when the legal conclusion is correct. But there is a risk in this framing that should be registered. Pacific values and social networks should not be invoked to excuse favouritism, silence complainants or avoid enforcing the law. Cultural understanding must sharpen accountability, not dilute it.
Large jurisdictions can move a matter to another office or region when conflicts arise. Small ones often cannot. This is why pre-agreed external assistance pathways matter, covering legal authority, information sharing, confidentiality, cost and who owns the decision. They should be settled before a difficult matter arrives, because arrangements improvised under pressure rarely inspire confidence. Regional networks, secondments and referral arrangements make independence more achievable.
Three further design choices follow. Conflict checks should run at intake, allocation, key decision points and closure, because circumstances change as a matter develops. Critical decisions should be separated even where whole units cannot be: the officer who receives a complaint need not be the one who decides whether it proceeds. And agencies should measure more than outputs. Complaint and investigation totals show activity, not whether independence is becoming credible. Recorded conflict checks, timely control decisions and confidence in reporting channels tell you more.
Independence requires protection from improper influence, but legitimacy also requires connection, explanation and accountability. An agency too exposed to relationships may be compromised; one that treats independence as social distance becomes isolated and poorly informed. The same tension faces regulators, auditors and public service commissions across the region.
In societies where everyone knows everyone, independence cannot rest on the fiction that public officials stand outside relationships. Nor should it accept the opposite fiction that relationships are harmless because they are normal. The practical test is whether an institution can show, decision by decision, that connections were disclosed, risks were managed, reasons were recorded and relationships did not determine the outcome. Legal establishment creates the institution; practised independence enables it. How these disciplines work in different Pacific settings is one of the questions on the region’s anti-corruption research agenda.
