Who Should Control Public Research Funding: Experts, Citizens, or Politicians?

Getting your Trinity Audio player ready...

Public research funding should not be controlled exclusively by experts, citizens, or politicians. Politicians should determine the total public budget and broad national priorities; citizens should help identify social needs and evaluate public impact; experts should assess scientific validity; and independent institutions should make individual funding decisions under transparent rules.

This division of authority is necessary because research funding involves several different questions:

• How much public money should be spent on research?
• Which social problems deserve priority?
• Which proposed methods are scientifically credible?
• Which researchers or outputs merit funding?
• How should funding decisions be audited?

No single group is qualified to answer all five.

Giving complete control to experts risks professional insularity and conflicts of interest. Giving it to citizens risks popularity replacing scientific judgment. Giving it to politicians risks short-term political interests distorting long-term research. A robust system separates these functions instead of searching for one ideal decision-maker.

Why Public Research Funding Requires Democratic Control

Public research funding is financed by taxpayers and affects the future direction of society. Decisions about its overall scale and purpose are therefore political in the legitimate sense of the word.

Elected governments must be able to decide whether additional resources should go toward science, education, healthcare, infrastructure, defence, welfare, or debt reduction. Scientists cannot unilaterally determine how much of the public budget they should receive.

Governments also have a legitimate role in establishing broad research missions. A democratic society may decide to prioritize areas such as:

• cancer prevention;
• pandemic preparedness;
• clean energy;
• food security;
• cybersecurity;
• fundamental mathematics;
• climate adaptation;
• strategically important technologies.

The European Union’s Horizon Europe Missions, for example, organize research around large, time-bound societal goals involving researchers, policymakers, citizens, and other stakeholders.

However, democratic authority over priorities does not imply that politicians should select individual researchers or determine which scientific claims are correct.

A minister can legitimately say, “The country should invest more in antibiotic resistance.” The minister should not ordinarily decide, “Professor A’s molecular mechanism is more plausible than Professor B’s.”

That second decision requires specialized scientific evaluation.

What Experts Are Qualified to Decide

Experts are indispensable because scientific proposals cannot be evaluated through popularity alone.

A technically competent assessment may require knowledge of:

• existing research literature;
• experimental design;
• mathematical validity;
• statistical power;
• laboratory feasibility;
• data quality;
• ethical limitations;
• replication risks;
• the difference between an incremental and a genuinely original result.

The US National Science Foundation evaluates proposals using two principal criteria: intellectual merit and broader impacts. Its process combines assessment of scientific quality with consideration of potential social benefit. The NSF merit-review system illustrates why expert evaluation remains central even when public impact is part of the decision.

Similarly, the US National Institutes of Health uses scientific review groups composed primarily of non-federal scientists, followed by advisory review that includes scientific and public representatives. The NIH grant review process is therefore not purely technocratic: it separates detailed scientific assessment from broader institutional funding judgment.

Experts should generally control questions such as:

• Is the research method valid?
• Does the proposal engage seriously with existing knowledge?
• Are the conclusions supported by the evidence?
• Is the proposed work technically feasible?
• Does the result constitute a meaningful scientific contribution?
• Can another researcher reproduce or verify it?

Citizens and politicians can define desired outcomes, but they cannot settle a difficult theorem, assess a novel particle detector, or determine whether a biomedical study is statistically sound without expert assistance.

Why Expert Control Alone Is Insufficient

Peer review is necessary, but it is not infallible.

Experts may have conflicts of interest. They may favour established institutions, familiar methodologies, fashionable subjects, or researchers within their professional networks. Highly original work may be difficult to evaluate because few reviewers understand it. Entire disciplines can develop shared blind spots.

Expert-controlled funding may also reproduce the Matthew effect: researchers who already possess funding, reputation, institutional support, and collaborators become more likely to receive further funding.

Even official funding bodies acknowledge these risks. NIH introduced a simplified review framework partly to reduce the influence of reputational bias and focus attention more directly on scientific merit. Its current peer-review framework reflects the continuing need to improve expert evaluation rather than treat it as automatically objective.

Peer review can also become conservative. Reviewers are often more comfortable evaluating projects that extend known research programmes than proposals based on unfamiliar concepts.

This produces a structural problem:

Experts are best positioned to evaluate technical quality, but existing experts may also be professionally invested in the theories, institutions, and funding patterns being challenged.

The solution is not to eliminate experts. It is to make expert judgment plural, contestable, documented, and auditable.

What Citizens Should Control

Citizens possess a different kind of knowledge from scientific specialists.

Patients understand the practical burden of diseases. Residents understand local environmental hazards. Farmers understand agricultural constraints. Disabled people understand which technologies address real accessibility problems. Communities living near industrial sites may recognize risks that are invisible in national statistics.

Citizen participation can therefore improve research agenda-setting by identifying neglected needs.

The European Commission describes citizen engagement as participation in the co-design of research agendas, co-creation of research, and co-assessment of outcomes. Its Horizon Europe citizen-engagement framework argues that participation can align research with social needs and improve trust and impact.

Citizens can contribute meaningfully to decisions such as:

• Which health conditions are being neglected?
• Which environmental outcomes matter locally?
• What risks are socially acceptable?
• Which research results should be made publicly accessible?
• Are funded projects producing benefits that people can actually use?
• Which practical problems deserve dedicated funding programmes?

Citizen assemblies, consultations, patient organizations, civil-society panels, participatory budgeting, and public-interest representatives can all contribute.

But citizen participation should not be confused with voting directly on the technical merit of every proposal.

Why Popular Voting Cannot Replace Scientific Review

A public vote rewards visibility, emotional appeal, communication skill, and existing popularity. Those qualities may be valuable, but they are not identical to scientific importance.

A campaign to develop a visible medical device may attract more votes than foundational work on an obscure mathematical structure, even when the mathematical work could eventually transform several scientific fields.

Likewise, rare diseases, unglamorous infrastructure, replication studies, dataset maintenance, and negative results may receive little public attention despite having substantial scientific value.

Direct voting also creates opportunities for manipulation:

• wealthy organizations can purchase publicity;
• charismatic personalities can outperform careful researchers;
• simplified claims can defeat technically accurate explanations;
• coordinated groups can dominate low-participation votes;
• immediate emotional concerns can crowd out long-term basic research.

Citizen participation is most effective when it shapes values, priorities, needs, and accountability, not when it substitutes mass opinion for technical evaluation.

The Proper Role of Politicians

Politicians should exercise strategic and fiscal authority, but their control should stop before individual scientific adjudication.

Legitimate political responsibilities include:

1. Setting the total research budget.
2. Balancing research against other public expenditures.
3. Defining broad missions and public-interest constraints.
4. Establishing laws concerning safety, ethics, transparency, and equal access.
5. Creating and supervising funding institutions.
6. Auditing whether public money is being used lawfully and effectively.

Politicians should generally not:

• choose grant recipients personally;
• alter scientific scores to reward allies;
• suppress lawful research because its conclusions are inconvenient;
• demand predetermined scientific findings;
• punish individual researchers for politically unpopular results;
• convert public research programmes into patronage systems.

The distinction is comparable to judicial independence. Legislatures create laws and budgets, but they should not ordinarily dictate the verdict in an individual case. Similarly, governments may establish research policy without deciding the scientific merit of every individual output.

A Better Model: Separation of Funding Powers

Research governance should use a constitutional structure rather than a single controlling authority.

Politicians: budgets and broad missions

Elected representatives authorize taxation, determine total expenditure, and define broad strategic objectives.

Their decisions should be public, time-limited, and subject to democratic review.

Citizens: needs, values, and impact

Citizens and affected communities help identify neglected problems, participate in mission design, and evaluate whether funded research serves the public.

Participation should be structured to include informed deliberation rather than relying only on online voting.

Experts: scientific evaluation

Independent experts assess methodology, originality, feasibility, evidence, and scientific contribution.

Reviews should include explicit reasoning, conflict-of-interest disclosure, and opportunities for challenge or additional review.

Independent funding institutions: implementation

Operational institutions should combine these inputs according to published rules. Their mandate should protect individual decisions from arbitrary political intervention while preserving public accountability.

Auditors: procedural oversight

Independent auditors should examine:

• conflicts of interest;
• unexplained deviations from evaluation criteria;
• concentration of funding;
• discrimination between institutions or regions;
• compliance with open-data and reporting requirements;
• whether promised outputs were produced;
• whether evaluation procedures are being manipulated.

This structure prevents any one group from converting legitimate authority in one domain into unlimited power over the entire system.

Should Experts Determine Research Priorities?

Experts should advise on priorities, but they should not possess an exclusive veto over society’s objectives.

Scientists can explain:

• which threats are technically serious;
• which fields are underfunded;
• what research capacity already exists;
• which goals are feasible;
• what unintended consequences a programme may produce.

Citizens and elected representatives must then weigh that advice against social values and competing public needs.

For example, epidemiologists can estimate the potential benefit of pandemic-preparedness research. Economists can calculate opportunity costs. Citizens can express acceptable trade-offs. Politicians must ultimately authorize spending.

Priority-setting is therefore a mixed epistemic and democratic decision. It combines facts about what is possible with values about what society should pursue.

Basic Research Needs Special Protection

A purely mission-driven system would underfund research whose future applications cannot yet be predicted.

Fundamental mathematics, theoretical physics, evolutionary biology, astronomy, and other basic sciences often produce value over long and uncertain time horizons. Their importance may not be visible to voters, politicians, or even neighbouring disciplines.

A well-designed public funding portfolio should therefore reserve substantial resources for:

• investigator-initiated research;
• curiosity-driven research;
• high-risk theoretical work;
• preservation of scientific infrastructure;
• replication and validation;
• small contributions that improve shared knowledge;
• unexpected discoveries outside current political missions.

Experts should have greater autonomy within this protected basic-research allocation, although their decisions must still be transparent and reviewable.

Can Algorithms Improve Public Research Funding?

Computational systems can help combine democratic objectives with expert evaluation, but they should not become unaccountable rulers.

An AI-assisted funding system could:

• compare claims with the scientific literature;
• identify dependencies between research outputs;
• detect possible conflicts of interest;
• distribute evaluation across multiple reviewers;
• explain how each criterion affected a decision;
• track later evidence about whether a result remained useful;
• reward data, software, proofs, replications, and corrections separately;
• detect suspicious concentration of funding or recognition.

The AI Internet Meritocracy model proposes a more continuous and output-oriented approach. Instead of relying only on occasional grant competitions, a system could evaluate completed scientific contributions and allocate rewards according to their demonstrated merit and relationships to other work.

This does not remove the need for governance. The funding source must still determine:

• the available budget;
• the eligible fields and outputs;
• the public objectives;
• the evaluation rules;
• the appeal process;
• the limits placed on automated decisions.

AI can administer rules, aggregate evidence, and expose inconsistencies. It cannot legitimately decide society’s ultimate values by itself.

Why Transparency Matters More Than the Identity of the Decision-Maker

The question “Who controls research funding?” can conceal a more important issue: Can anyone understand, challenge, and audit the decision?

An expert committee operating through opaque deliberations can behave politically. A citizen panel can be captured by activists. A ministry can rely on excellent independent advice. An algorithm can either expose every relevant factor or conceal discrimination behind a numerical score.

Good governance therefore requires:

• published criteria;
• written explanations;
• disclosed conflicts of interest;
• access to evaluation evidence;
• identifiable responsibility for decisions;
• appeal or reconsideration procedures;
• protection against retaliation;
• publication of aggregate funding patterns;
• independent audits;
• periodic evaluation of the funding system itself.

Every rejection should receive an intelligible explanation. Every award should be traceable to stated criteria. When the system departs from expert rankings for strategic or social reasons, that departure should be explicit rather than hidden.

Transparency does not eliminate disagreement, but it makes disagreement productive. Researchers can correct errors, citizens can identify neglected needs, politicians can evaluate programme performance, and auditors can detect capture.

A Practical Governance Architecture

A balanced public research fund could operate in five stages.

1. Democratic budget authorization

The legislature determines the total budget, legal constraints, and broad distribution between fundamental research, strategic missions, infrastructure, and other categories.

2. Participatory priority-setting

Citizens, civil-society organizations, researchers, patients, industries, and affected communities contribute to the design of research missions.

This process identifies problems rather than selecting predetermined scientific conclusions.

3. Independent scientific assessment

Qualified reviewers evaluate technical merit. Multiple reviewers, structured criteria, conflict controls, and transparent reasoning reduce the influence of individual bias.

4. Rule-based funding allocation

An independent agency or auditable computational mechanism combines scientific assessments, public priorities, portfolio requirements, and available resources.

For completed research outputs, retroactive or continuous rewards can supplement traditional prospective grants. This is one reason to examine alternative models of scientific funding rather than treating advance grants as the only possible mechanism.

5. Public audit and revision

Results, evaluations, funding concentrations, appeals, and programme outcomes are reviewed publicly. The rules can then be amended without allowing politicians to rewrite individual scientific judgments retroactively.

Who Should Have the Final Word?

There should not be one final decision-maker for every type of question.

Decision	Primary authority	Necessary checks
Total research budget	Politicians	Elections, fiscal transparency
Broad social priorities	Politicians and citizens	Expert feasibility advice
Scientific validity	Experts	Multiple review, disclosure, appeals
Individual funding allocation	Independent institutions	Published rules and audits
Evaluation of social impact	Citizens, experts, and public bodies	Evidence and open reporting
Ethical and legal boundaries	Democratic institutions	Scientific and civil-society consultation
Performance of the funding system	Independent auditors and the public	Accessible data and periodic review

The governing principle is simple:

Democracy should decide what public research is for, expertise should determine what is scientifically credible, and independent transparent systems should determine how general rules apply to individual cases.

Conclusion

Experts, citizens, and politicians each possess legitimate authority, but in different domains.

Politicians control public budgets because taxation and national priorities require democratic authorization. Citizens should influence the problems research addresses because they experience its consequences and finance it. Experts must evaluate scientific merit because technical validity cannot be decided through elections or popularity. Independent institutions must convert these inputs into individual funding decisions without becoming instruments of political patronage or professional self-interest.

The strongest model is therefore not expert rule, direct democracy, or ministerial control. It is a separation of funding powers supported by transparency, plural evaluation, public participation, independent administration, and continuous audit.

Public research funding should be democratically directed, scientifically evaluated, and institutionally protected.