Can governance protect a decision without weakening its evidence?
The expertise-independence trade-off
Introduction
Conflict-of-interest governance usually begins with a sensible objective: protect a decision from influence that should not determine its outcome. Depending on the circumstances, that may require disclosure, recusal, restricted access to information, reassignment, independent review, additional oversight or another proportionate control.
The difficulty is that expertise and independence are not always distributed neatly across different people. The executive who helped design a commercial arrangement may understand its economics better than anyone else. A technical specialist may possess knowledge that nobody else holds while also having a professional interest in one solution. An HR leader who helped design a remuneration proposal may later be needed to explain the assumptions behind it. A founder whose involvement creates concentration risk may also hold knowledge that has never been fully institutionalised.
This creates a harder governance question. Removing inappropriate influence can strengthen independence, but it can also remove decision-relevant information. A process can therefore become more formally independent while simultaneously becoming less informed.
The central question is not simply whether a conflicted person has been properly constrained. It is whether the organisation has protected independence without degrading the evidence base on which the decision depends.
That is the expertise-independence trade-off at the centre of this edition.
Research and validation approach
This edition draws on four complementary evidence streams. The first is conflict-of-interest guidance from the OECD and the UK Government. These sources make clear that declaration alone does not resolve a conflict and that appropriate mitigation can involve different forms of restriction, reassignment, oversight and monitoring rather than only complete exclusion.
The second stream is corporate-governance guidance from the UK Financial Reporting Council. The UK Corporate Governance Code emphasises objective judgement, constructive challenge, appropriate expertise and access to accurate, timely and clear information. This is important because governance quality depends not only on who is independent, but also on whether decision-makers receive evidence of sufficient quality to exercise that independence meaningfully.
The third stream comes from institutional practices used by the US Food and Drug Administration and the National Academies. These are highly specific governance regimes and should not be transplanted mechanically into corporate settings. However, they demonstrate an important principle: there can be circumstances in which independence concerns and access to scarce expertise must be managed simultaneously rather than treated as entirely separate questions.
The fourth stream comes from research on group decision-making. The hidden-profile literature shows that groups can collectively possess the information needed for a strong decision while still failing to surface or use information held uniquely by individual members. This matters because removing a person from a decision process can also remove information that nobody else possesses.
Taken together, the evidence does not support the conclusion that conflicted experts should remain influential simply because they are knowledgeable. Nor does it support the opposite conclusion that greater exclusion will always improve decision quality. The real governance task is to separate expert contribution from decision influence.
Evidence context 1: Conflict control is not always binary
Many organisations instinctively treat conflict management as a binary choice: participate or recuse, remain involved or leave the process, vote or do not vote, retain access or lose access. In practice, mature governance can be more granular.
OECD guidance recognises several possible responses, including recusal and restriction of information, and cautions that excessively rigid controls can become impractical or counterproductive. UK procurement guidance similarly describes a range of mitigation mechanisms, including reassignment, multiple assessors, independent observers, management review and continuing monitoring.
The FDA provides a different example from a regulated scientific environment. Advisory-committee participation by individuals with relevant financial interests is tightly controlled, but the rules recognise limited circumstances in which the need for an individual’s services may outweigh the potential conflict, subject to formal safeguards. Participation can also be restricted, for example by removing voting authority.
The National Academies applies similarly demanding independence standards while recognising that, in limited circumstances, equivalent non-conflicted expertise may not be available. Again, these regimes are not direct templates for corporate governance. Their importance lies in the principle they illustrate.
Access to expertise and authority over the decision do not always have to reside in the same person.
That distinction allows governance to become more precise. The objective is not necessarily to decide whether someone is simply “in” or “out”. It is to determine what contribution is necessary, what influence is inappropriate and how the two can be separated.
Evidence context 2: Removing a person can also remove unique information
Group-decision research introduces a second risk. A committee may collectively possess the information required for a strong decision and still fail to use it effectively.
A major meta-analysis of hidden-profile research covered 65 studies, 101 independent effects and 3,189 groups. It found that groups discussed substantially more shared information than uniquely held information, and that stronger pooling of unique information was positively associated with better decision quality.
The organisational implication is important because expertise is rarely distributed evenly. One risk specialist may understand a regulatory dependency that nobody else has identified. One commercial leader may know why an apparently attractive option cannot be implemented. One HR professional may understand why a market comparison is structurally misleading. One technical specialist may possess historical context that is absent from formal documentation.
If governance removes such a person without first preserving the decision-relevant knowledge they hold, formal independence may improve while substantive decision quality deteriorates.
The problem has not disappeared. It has changed form.
Instead of excessive influence, the organisation may have created evidence loss.
The critical conceptual distinction: Contribution is not the same as control
A conflicted or insufficiently independent expert may still possess valid information. The governance problem arises when access to expertise also gives that person disproportionate influence over how the decision is constructed.
Influence can occur long before the final vote. It can appear in the definition of the question, the selection of evaluation criteria, the evidence admitted into the process, the framing of alternatives, the assumptions treated as settled, the description of uncertainty, the drafting of recommendations or the scoring of options.
This means that a person can materially shape a decision without formally approving it.
A stronger governance design therefore separates legitimate contribution from decision control. An expert may provide factual evidence without determining the evaluation criteria. They may explain technical constraints without scoring the alternatives. They may clarify assumptions without drafting the final recommendation. They may answer questions without participating in final deliberation. Their evidence may remain available while being independently tested by others.
The important design question is therefore not simply whether the person should participate. It is which parts of the decision genuinely require their knowledge and which parts require protection from their influence.
That is a much more demanding question than recusal alone.
Further organisational interpretation
The problem becomes especially important when expertise is highly concentrated.
An organisation may believe it has a conflict-management problem when it also has a capability-distribution problem. If one person cannot be removed from a consequential decision without materially weakening the organisation’s ability to understand the issue, the conflict may be exposing something deeper: decision-critical knowledge has become too concentrated.
That concentration may reflect limited succession depth, weak knowledge transfer, poor documentation, concentrated stakeholder relationships, insufficient technical bench strength, an informal operating model or limited independent challenge capability.
The conflict does not necessarily create these weaknesses. It reveals them.
This produces a second-order governance question. Why can the organisation not protect independence without simultaneously losing decision capability?
That question matters because a business that cannot remove one individual from a decision without becoming materially less capable is carrying a form of organisational dependency, regardless of whether the individual is conflicted.
In that sense, conflict governance can become an unexpected test of organisational resilience.
The RewardEdge expertise preservation test
The RewardEdge Expertise Preservation Test is a practitioner framework rather than an established academic, legal or regulatory instrument. Its purpose is to help leaders examine the quality of the decision architecture when an individual’s independence is in question but their expertise remains potentially relevant.
1. Decision-relevant expertise
The first question is what specific knowledge the individual possesses that could materially affect the quality of the decision. Vague claims such as seniority, experience or organisational memory are insufficient. The organisation should identify the actual evidence, judgement or context that could be lost.
2. Substitutability
The second question is whether equivalent expertise can reasonably be obtained from another credible and sufficiently independent source. If it can, substitution may provide the cleaner governance solution. If it cannot, the participation design requires greater care.
3. Contribution boundary
The third question is what the individual may legitimately contribute. This might include factual evidence, technical explanation, historical context, implementation constraints or scenario analysis. The boundary should be explicit rather than assumed.
4. Influence boundary
The fourth question is which parts of the decision must remain outside that person’s control. Depending on the situation, this may include evaluation criteria, option framing, scoring, recommendation drafting, deliberation, voting or final approval.
5. Independent testing
The fifth question concerns challenge. Who independently tests or validates the expert contribution? Scarce expertise should not become self-authenticating simply because few people possess equivalent knowledge.
6. Deliberation protection
The sixth question is whether accountable decision-makers can assess the evidence, discuss alternatives and reach a conclusion without the expert dominating the deliberation. A process can preserve access to knowledge while still protecting the independence of judgement.
7. Capability concentration
The final question is whether removing the individual would leave the organisation materially less capable of reaching a competent decision. If the answer is yes, that should be treated as a separate organisational risk rather than simply as a reason to tolerate the conflict.
Taken together, the framework moves beyond the simple sequence of conflict followed by recusal. A more complete sequence is conflict, expertise need, substitutability, contribution boundary, influence boundary, independent testing, protected deliberation and final decision.
Important counterpoint
There is an obvious danger in this argument. “We need their expertise” can easily become an excuse for allowing a conflicted individual to retain too much influence.
That would defeat the purpose of the governance control.
Scarce expertise does not prove that participation is safe. Seniority does not create an exemption. Technical reputation does not eliminate bias risk. Long tenure does not entitle someone to control a decision. Concentrated knowledge should not automatically translate into expanded decision authority.
Where reasonably equivalent independent expertise is available, using it may be the more defensible approach. Where the conflict is sufficiently serious, protection of the process may justify excluding the individual even when some information is lost. Where participation is retained, the boundaries should be explicit, documented and independently governed.
The objective is therefore not to make conflict management more permissive.
It is to make conflict management more precise.
The hidden Reward, job architecture and workforce implication
There is also a Reward implication, but Reward is not the first solution.
When an organisation repeatedly discovers that one executive or specialist is indispensable to critical decisions, it may respond through retention payments, market premiums, larger roles, special allowances, accelerated progression or other differentiated Reward mechanisms.
Some of those responses may be justified. Scarce capability can have genuine economic value.
The risk arises when compensation begins to conceal an operating-model problem.
A stronger sequence is to identify why the decision-critical expertise is concentrated, determine whether that concentration is deliberate and economically justified, strengthen documentation and capability transfer where appropriate, assess whether enduring accountability or scarcity genuinely belongs in the job, and only then determine whether differentiated pay is justified.
Otherwise, the organisation may end up paying a premium for dependency that its own operating model created.
This is adjacent to previous RewardEdge work on capability dependency, but the unit of analysis is different. The question here is not simply whether scarce capability creates retention risk. It is whether the organisation can maintain decision independence when decision-critical expertise is concentrated.
What the evidence supports
The evidence supports several bounded conclusions.
Conflict declaration alone is not sufficient conflict management. Conflict controls can take different forms rather than being limited to complete exclusion. Effective governance depends not only on independence but also on access to high-quality, timely and challengeable information. Group decision-making research shows that uniquely held information can be underused, and that better pooling of such information is associated with better decision quality. Certain institutional regimes also recognise limited circumstances in which the need for expertise and concerns about independence must be managed together through explicit safeguards.
These findings support the central argument of this edition: organisations should distinguish the value of expert information from the authority to shape or determine the decision.
What the evidence does not support
The evidence does not establish that expertise should override independence. It does not support routine participation by conflicted experts. It does not show that partial participation is appropriate in every legal, regulatory or organisational setting. It does not prove that every conflict can be solved through role separation, nor that the loss of expertise is always more serious than the conflict itself.
The RewardEdge Expertise Preservation Test should therefore be understood as a practitioner synthesis rather than an academically validated instrument. Its purpose is to improve the questions leaders ask. It does not replace applicable legal, regulatory, ethical or professional requirements.
Five common barriers
The first barrier is status. Senior specialists may interpret participation limits as distrust or loss of standing, particularly where their expertise has previously given them broad influence over a decision domain.
The second barrier is scarcity. Organisations often discover too late that critical knowledge has never been adequately transferred, documented or challenged by others.
The third barrier is process ambiguity. A person may be described as recused without anybody defining whether that means no vote, no recommendation, no deliberation, no access to papers or no contribution at all. Without that clarity, recusal may exist only as a label.
The fourth barrier is informal influence. Someone excluded from the formal decision meeting may already have shaped the assumptions, criteria, available options or recommendation through earlier involvement.
The fifth barrier is false independence. Replacing a conflicted expert with somebody who is formally independent but materially less informed can improve the appearance of governance without necessarily improving the quality of the decision.
Independence matters. Competence matters as well. Strong governance has to protect both.
Practical strategies
A practical response begins by identifying exactly which decision-relevant information may be lost if participation is restricted. Evidence submission can then be separated from recommendation authority, and specialist assumptions can be documented so that they are visible and challengeable.
Where proportionate and feasible, independent technical review can provide an additional safeguard. Organisations can also separate fact-finding, option development, evaluation, recommendation and approval when the significance of the decision justifies stronger role separation.
The source of material evidence should be documented, and the extent of an individual’s permitted involvement should be clear. Decision-makers should know whether that person may provide information, answer questions, join deliberation, score options or vote.
Finally, once the immediate decision is complete, the organisation should ask whether its dependence on the expert has revealed a broader succession, documentation or capability-transfer problem. Conflict management should not end when the meeting ends if the process has exposed a deeper organisational dependency.
Implementation in the operating model
This issue should not sit only inside Legal or Compliance. It belongs in the operating model because it concerns how evidence, expertise, challenge, recommendation and authority move through the organisation.
For sufficiently consequential decisions, organisations can distinguish among the roles of evidence provider, subject-matter expert, technical challenger, decision sponsor, recommendation owner, independent reviewer and final decision authority.
The same person may legitimately hold more than one role where governance allows. However, where independence is material, role separation should be deliberate rather than assumed.
The objective is not to bureaucratise every organisational decision. It is to apply stronger separation where the combination of consequence, conflict risk and concentrated expertise makes it necessary.
Five leadership actions
1. Map expertise before removing influence
Leaders should understand what decision-relevant knowledge may disappear before determining the appropriate control. Removing influence without understanding what information will also be removed can create a different form of decision risk.
2. Separate knowledge from authority
Access to expertise should not automatically require access to decision control. Organisations should deliberately distinguish the right to inform a decision from the right to frame, recommend or approve it.
3. Protect the deliberation
Independent decision-makers need sufficient space and authority to test evidence, challenge assumptions and reach their own conclusion. Independence is weakened if the expert remains able to dominate how the evidence is interpreted.
4. Test for expertise concentration
If removing one person materially damages the organisation’s ability to decide competently, that should be treated as a separate organisational dependency requiring attention beyond the immediate conflict.
5. Review both integrity and decision quality
A process can be formally independent and still be poorly informed. Good governance should therefore protect both the integrity of the process and the quality of the evidence used within it.
The RewardEdge perspective
Governance is sometimes discussed as though independence improves automatically as participation decreases. That is too simple.
A decision can be weakened by excessive influence. It can also be weakened by the loss of information that nobody else possesses. The objective should therefore be neither maximum participation nor maximum exclusion.
The stronger objective is sufficient expertise to inform the decision without allowing the source of that expertise to inappropriately determine the outcome.
That requires organisations to distinguish contribution from control, expertise from authority, information from influence and formal independence from actual decision quality.
The deeper warning is that when one person is too conflicted to control a decision but too indispensable to remove from it, the organisation may have two governance problems rather than one.
The first is conflict.
The second is concentrated capability.
And solving only the first may leave the second untouched.
Inviting perspectives
I would especially value perspectives from ten voices whose recent thinking intersects with different parts of this problem.
Wojciech Z. Kaleta, 🎓 PhD: Where should expert contribution end and decision authority begin when the expertise itself is scarce?
Andrew Nowak: When does protecting formal independence begin to weaken organisational judgement rather than strengthen it?
Chris Roberts: How should organisations preserve technical contribution when the expert may already have influenced the framing, criteria or alternatives?
Brian Jackson: Can organisations reliably separate expert contribution from decision constraint once that expertise has become embedded in the process?
Carl Mikael Björn: At what point should independent revalidation become necessary when material evidence originates from a conflicted source?
Eric Lint: How far should governance follow the influence when expert assumptions have already shaped downstream analysis or recommendations?
Catherine McDonald: What does organisational learning look like when a governance process reveals excessive dependence on one person’s judgement?
HILARY M CORNA: When does expert dependency signal that a process itself has failed to become repeatable and scalable?
Sergio D’Amico, CSSBB: How do organisations know whether they have transferred real judgement and problem-solving capability rather than merely documented the standard work?
Olaf Boettger: From a continuous-improvement perspective, when should repeated reliance on expert intervention trigger capability redesign rather than further dependence on the expert?
Closing question
If removing one individual protects the independence of a decision but materially weakens the organisation’s ability to understand that decision, has governance really been strengthened, or has one risk simply been exchanged for another?
Subscribe to RewardEdge Daily Insights for evidence-led practitioner analysis on Reward, work, organisation design, leadership, AI, governance and the workforce decisions that shape organisational value.
RewardEdge Consulting supports organisations reviewing Reward governance, job architecture, workforce design, decision architecture and organisational controls.
Sources and further reading
1. OECD. Managing Conflict of Interest in the Public Service: OECD Guidelines and Country Experiences. OECD Publishing, 2004.
2. UK Government. Guidance: Conflicts of Interest under the Procurement Act 2023. Updated 13 July 2026.
3. Financial Reporting Council. UK Corporate Governance Code 2024. Published 22 January 2024.
4. Financial Reporting Council. Corporate Governance Code Guidance.
5. US Food and Drug Administration. Public Availability of Advisory Committee Members’ Financial Interest Information and Waivers.
6. National Academies of Sciences, Engineering, and Medicine. Composition and Balance, Conflicts of Interest, and Independence: Policies and Procedures.
7. Lu, L., Yuan, Y. C. & McLeod, P. L. “Twenty-Five Years of Hidden Profiles in Group Decision Making: A Meta-Analysis.” Personality and Social Psychology Review, 2012, 16(1), 54–75. DOI: 10.1177/1088868311417243.
