Previous
This short series titled Evidence as Politics in Policy explores a simple but often overlooked point: evidence does matter in policymaking, but never on neutral ground. In all political systems, evidence is used strategically, selectively ignored, unevenly protected, and shaped by democratic and non-democratic pressures, institutional design, and political incentives. Across these articles, I argue that the challenge is not simply to produce more evidence or communicate it better, but to understand the political, economic and social conditions under which evidence is used, contested, or sidelined — and what that means for anyone interested in evidence-informed policy.
The reflections in these articles have been drawn from a series of interviews with political scientists in Latin America, Africa and Asia as part of our ongoing effort to better understand the political economy of evidence use.
This last article provides a summary of notes from 6 conversations, plus additional reading of the experts’ published work.
Six experts, six ways of understanding why evidence matters less than we like to think
I did not have time to review every interview and relevant additional sources in detail for this series but this bonus article captures the views from the conversations.
We often speak about evidence-informed policymaking as if the main challenge were better data, clearer communication, or more rigorous methods. But the experts referenced in this article — Eduardo Dargent, Adolfo Garcé, Jorge Mori, Tricia Yeoh, Emmanuel Remi Aiyede and Frederick Golooba-Mutebi— and whose insights inform the series point in a different direction. Their arguments suggest that the real question is not whether evidence exists, but what political, institutional and intellectual conditions allow it to matter at all.
None of them treats evidence as a neutral input that naturally improves policy. All, in different ways, show that evidence is filtered through power, ideas, institutions, leadership, incentives, and conflict. Evidence does not enter policy on its own; it has to be carried, defended, translated, and made useful in political contexts that are often indifferent or openly hostile to it.
After this synthesis, I have included 6 short Q&As based on the interviews (with the help of AI).
1. Evidence is political before it is technical
The strongest point of agreement among all five experts is that evidence is always political.
Eduardo Dargent is especially clear on this. The main barriers to evidence-informed policy are not primarily methodological or informational. They lie in power dynamics, veto actors, and the ability of organised interests to bend policy to their advantage. In Technocracy and Democracy in Latin America, Dargent asks why technocrats retain influence in some sectors and not in others, already shifting the conversation away from evidence as a universal good and toward the political conditions that protect or weaken expertise.
His work with Gabriela Spanghero Lotta pushes that argument further by treating experts as political actors whose relationship with politicians is often tense, conflictive and contingent rather than neutral or harmonious.
Keep reading+
Frederick Golooba-Mutebi arrives at a similar conclusion, but from the perspective of implementation. He argues that politics is not just a blockage for evidence use, but it often changes the timing of policy action so drastically that the evidence on which a policy was originally designed is no longer current by the time governments actually implement it. Plans are drawn up, delayed by political imperatives, then rushed forward near elections or in response to public pressure. By then, the facts on the ground may have shifted. Evidence, in other words, can be made obsolete by the rhythms of politics. To this he adds another, more material constraint: governments and universities often do not invest enough in local research, while much externally produced research remains difficult to access.
Goloomba-Mutebi’s views are strongly influenced by his work on political settlements in Africa.
Keep reading+
Jorge Mori reaches a conclusion similar to Dargent’s in a more openly strategic language: evidence is a weapon or a tool, not a neutral object. While Dargent emphasises how power blocks or distorts evidence, Mori reminds us that evidence can itself be deployed in political struggle, either to advance reform or to defend the status quo.
Mori’s work on Peruvian university reform is concerned with reform as a contested process rather than a settled technical fix, while his later work on graduate tracking systems points to the need for durable state capabilities and ongoing information systems, not just isolated studies.
Keep reading+
Tricia Yeoh also treats evidence as inseparable from politics, but her emphasis falls less on veto players and more on political acceptability. Evidence is taken up when there is an intermediary capable of presenting it in a form the government can use, and when politicians see reputational or strategic benefits in adopting it.
Her work on Malaysian federalism and state-federal relations is attentive to institutional design, but also to the negotiations, alignments and pressures that shape whether governments move at all. Her OTT keynote on bottom-up accountability adds another layer: policymakers often respond well when evidence is embedded in wider public or civic pressure, not when it simply arrives as a technical recommendation.
Keep reading+
What distinguishes Adolfo Garcé is that he pushes the analysis one level deeper. He does not stop at power. He asks about the ideas and traditions that make some forms of evidence more legible and legitimate than others.
That is consistent with his broader body of work on policy ideas, ideational diffusion, and political knowledge regimes. In other words, before evidence can be used, it must first fit within a political and intellectual framework that tells actors what counts as a problem, what kinds of knowledge are valid, and who is entitled to speak.
Keep reading+
2. The tension between technocracy and democracy
A second theme that cuts across these discussions is the uneasy relationship between expert knowledge and democratic politics.
Here, Garcé contrasts countries with strong technocratic or positivist traditions with those whose political life has been more pluralistic, democratic, or participatory. He suggests that the more democratic and participatory a system becomes, the harder it may be for technical expertise to move smoothly into public policy. That is not, for him, a reason to lament democracy but a reminder that democracy requires attention to the knowledge of citizens, workers, parties and social organisations, not only to universities and think tanks.
His work on policy ideas and political knowledge regimes helps situate this claim within a larger concern: institutions do not merely process information; they embody prior assumptions about authority, legitimacy and knowledge.
Golooba-Mutebi reframes a similar problem in more overtly political-system terms. His distinction is not between technocratic and participatory traditions as such, but between competitive, adversarial, winner-takes-all systems and more consensus-based systems. In the former, politicians begin planning for the next election almost as soon as they take office, which pulls them toward short-term decisions and away from long-horizon planning. In the latter, where ruling parties are more confident about political continuity and where multiple parties are already folded into governance arrangements, longer-term planning becomes more feasible. This does not amount to a defence of any particular regime model. But it does underline a point the evidence field often avoids: the time horizons built into political systems matter enormously for whether evidence can be used consistently.
Mori shares the concern about exclusion but frames it differently. His argument is that the use of evidence has become too elitist, dominated by experts and technocrats, especially economists, and needs to be democratised. His point is not that expertise should disappear, but that a healthy evidence ecosystem must widen who can use and interpret evidence, including teachers and other actors too often excluded from policy debates.
His practical work in higher education reform and graduate follow-up systems reinforces this concern with institutional inclusion and sustained use rather than one-off technical interventions by experts.
Yeoh also points to this tension. Public opinion polls count as evidence, even if they can be manipulated politically. They broaden what enters the conversation. Her work suggests that evidence is not simply technocratic analysis handed down by experts but rather something mediated by coalition politics, public legitimacy, and the difficult negotiation between state capacity and democratic responsiveness.
Dargent is less interested in theorising democracy-versus-technocracy as such than in describing what happens when political actors actively undermine technical sectors. Polarisation re-politicises settled issues, installs unqualified personnel, and intimidates bureaucrats into silence.
Aiyede adds another layer. Even where evidence exists and leaders are nominally interested in it, bureaucratic routines and administrative inertia can still slow or block change. The problem, then, is broader than a simple opposition between technocracy and participation. Evidence must pass through real organisations, with all their constraints, habits and incentives – the mediation referred to by Tricia Yeoh.
3. Leadership and intermediaries matter more than evidence advocates often admit
A third point of overlap is the importance of people — not just systems, not just data, but actual individuals occupying key positions.
For Garcé, leadership profiles are decisive. He argues that well-educated political leaders, especially those trained in public policy or related fields, are more likely to seek and use diverse forms of evidence. Macro-level institutions matter over time, but specific leaders matter immediately. That is very much in keeping with Garcé’s broader interest in how ideas travel into institutions and how political actors mediate them.
Yeoh places similar weight on agency, but through intermediaries. Evidence is more likely to be adopted when a credible actor — a think tank, expert, consultant, or respected institution — can make it usable for the administration. That is a powerful insight for anyone working in the evidence field.
Mori’s notion of gatekeepers points in the same direction. Former policymakers in think tanks or private institutions can mediate between evidence and government, especially when the current administration is uninterested in or hostile to evidence.
Aiyede makes a similar point in his discussion of ministers, policy entrepreneurs, and actors able to convert research into decisions. His emphasis on communication is not superficial. It is really about the actors and platforms that make research travel.
Even Dargent, whose writing is more structural, implicitly agrees. His distinction between protected, professionalised sectors and politically captured ones only makes sense because leadership and institutional guardianship matter. Someone has to defend technical authority if it is to survive. And he has shown what happens when those defenders leave.
Golooba-Mutebi helps sharpen this further. In his account, leadership matters not just because leaders can demand evidence, but because leaders shape the political horizon within which institutions operate. A government confident of remaining in power can plan longer term, while a government worried about losing office is more likely to crowd out evidence with urgent political calculations. That is a useful reminder that “leadership support” for evidence is not only about personal conviction but more about whether leaders feel able to govern beyond the next electoral cycle.
4. Some sectors use evidence more consistently than others
The experts also agree that the use of evidence is not evenly distributed across the State.
Golooba-Mutebi introduced the concept of pockets of effectiveness. Even in adversarial and highly politicised systems, he argues, some agencies or departments may be insulated enough from political contestation to plan and implement effectively. His examples from Uganda suggest that these pockets can exist for a time when they are protected from wider political conflict, but that they can also be pulled back into those conflicts and lose their effectiveness. This is a useful corrective to blanket accounts of whole governments or whole sectors. It suggests that the use of evidence may survive, unevenly and temporarily, within protected institutional niches even when the broader system remains inhospitable.
Dargent similarly points to the Ministry of Economics and some social policy sectors as places where evidence persists because professionalisation and statistical capacity are stronger. He also distinguishes between stronger national bureaucracies and weaker regional or municipal governments. This aligns with the sectoral logic in Technocracy and Democracy in Latin America, which is explicitly interested in variation across policy areas rather than in broad claims about “the state” as a whole.
Yeoh offers a similar picture. Ministries linked to finance, investment and the economy use evidence more consistently because the cost of getting things wrong is high and scrutiny is intense. Education, by contrast, appears as unstable, symbolic and highly politicised. Her work reinforces the broader argument that institutional incentives and political arrangements vary sharply across sectors and levels of government.
That helps make sense of Mori’s bleak depiction of education as a war room. In his case, the point is not simply that education lacks data but that the sector is politically overloaded, contested, and vulnerable to abrupt shifts in direction. His work on university reform in Peru reinforces the sense that education reform is not a technical management problem but a terrain of dispute over regulation, standards, autonomy and control.
Aiyede adds a useful comparative twist by noting that some sectors and geographies rely more on evidence because pressure forces them to. Large, complex urban systems and digitally intensive sectors cannot rely on improvisation for long. In that sense, necessity can sometimes do what good governance rhetoric cannot. His work on policy action and research uptake helps formalise that insight by connecting evidence use to specific policy demands and stakeholder pressures rather than to abstract commitments to rationality.
5. International pressure has not disappeared, but it has changed
The most interesting difference among these experts may be how they see the role of international actors.
Dargent argues that the international environment has changed and that external leverage from development banks or U.S. or European actors has weakened. This matters because it suggests that one of the old supports exoskeleton for technical reform has faded, leaving local political actors freer to ignore evidence unless other pressures arise.
However, his broader work on expertise and politics does not rest on international pressure alone; it is entirely compatible with the view that expert authority is fragile if it depends on foreign institutional backers.
Yeoh suggests a somewhat different pattern. In more closed or authoritarian settings, external linkages may still compel governments to heed evidence, whether through trade obligations, international indices, or reputational concerns. Her work on Malaysian politics underscores how national political choices are shaped by wider institutional and reputational contexts.
Aiyede is even more direct: donor agencies, certainly in Africa, often matter more than local experts because they bring funding and can shape how programmes are designed and implemented. His work on research communication and on social protection policy in Nigeria supports that broader understanding of donor influence and policy uptake: evidence moves more easily when it is tied to resources, implementation pathways, and politically credible programmes.
Golooba-Mutebi fits partly with Yeoh and Aiyede here, but with a sharper distinction between externally pushed and internally driven reforms. In competitive systems, he suggests, evidence is more likely to be used when external development partners such as the World Bank push a policy and make financing contingent on evidence-based design or implementation. In those cases, governments may have less room to evade evidence requirements. But where policies are driven by internal political pressures, evidence is much less likely to be prioritised. This is a useful addition because it reminds us that not all donor influence works in the same way: sometimes it strengthens evidence use not by persuasion, but by changing the conditions under which governments can access money or proceed with implementation.
So the disagreement here is not over whether international actors matter, but how they matter and under what conditions. For Dargent, their old leverage has faded. For Yeoh and Aiyede, it mutates rather than disappears.
Evidence matters — but only when politics lets it
What these five experts offer, taken together, is not a theory of why evidence wins but a theory of why evidence so often does not.
- Eduardo Dargent reminds us that expertise operates inside systems of power and can be neutralised by veto actors or ideological capture.
- Adolfo Garcé insists that evidence always arrives inside prior traditions of thought — policy paradigms, beliefs, and political cultures that shape what counts as valid knowledge.
- Jorge Mori warns that evidence can become an elitist instrument unless its use is democratised and tied to reform coalitions.
- Tricia Yeoh shows that evidence needs intermediaries, political acceptability, and strategic value to be taken up.
- E. Remi Aiyede grounds the conversation in the difficult practical bridge between research production, bureaucratic routines, leadership, and policy demand.
- Frederick Golooba-Mutebi reminds us that evidence is also constrained by governments’ political time horizons, weak local research systems, limited access to knowledge, and the instability of implementation in competitive systems.
The lesson I take away from these accounts is that evidence does not enter policymaking because it is rigorous, true, or compelling on its own. It enters when institutions protect it, when leaders ask for it, when intermediaries translate it, when political incentives align with it, and when the underlying ideas of a system make room for it.
It echoes the argument I have made that think tanks are influential not because they have the power to be, but because other, more powerful actors find them useful.
That is a much richer language of “evidence-informed policy” than what is usually allowed. And it is also more realistic.