The Recognition Trap: Why the Global Knowledge Divide Persists

The divide in whose knowledge counts is structurally stable. Closing it is a coordination problem.

For more than a decade I have taught and done research in two worlds at once, in France and in Morocco. Much of that work has gone into building collaborations that connect a rising generation of African economists to established research institutions in France. From that position, one pattern is hard to miss.

In science, publication is how work earns its standing. Peer-reviewed journals are the mechanism through which a finding stops being one researcher’s claim and becomes knowledge that others cite, teach, build on, and carry into policy. Publishing in the leading journals of a discipline determines whose questions set the agenda and whose evidence reaches the rooms where decisions are made. For a researcher, being absent from those journals means being absent from the conversation.

That access is starkly unequal. A recent study of more than 450,000 articles in economics journals found that, comparing published articles with similar citation records, an article written solely from developing-country institutions has a 1.4 percent probability of appearing in a top-ten journal. For an article written solely from US institutions, the figure is 15.5 percent. The gap is roughly eleven to one, and it is not explained by the quality of the work. This is what I have come to call the recognition trap. The work meets the standard; what it struggles to obtain is recognition. And the two are linked, because recognition withheld long enough erodes the conditions for producing the work at all.

A trap with two locks

Why does the gap persist? Two mechanisms sustain it, and each reinforces the other.

The first is what I call the power-knowledge circuit. The countries and institutions that produce the analytical frameworks — that set the benchmarks and define what counts as development success — are doing more than analysis. Who produces the evidence produces the narrative, and command of the narrative is bargaining power in the negotiations that follow: financing terms, governance frameworks, the design of conditionality. The relationship also runs in reverse. The OECD benchmark is trusted because the institutions behind it have the leverage to make it the reference point. The exclusion of knowledge produced by African economists is both a consequence and a driver of this asymmetry. Power legitimates knowledge, and knowledge reproduces power.

The second is a voice-and-exit spiral. When researchers see, year after year, that their findings are ignored and their frameworks never adopted, the rational response is a quiet exit: publish less internationally, engage less with global debates, invest energy elsewhere. Exit thins the local evidence base. External frameworks fill the vacuum, and their uncritical application confirms the perception that local institutions cannot produce reliable knowledge. Marginalization produces exit, and exit deepens marginalization. The cost is not only local: talent is misallocated, research institutions are hollowed out, and global knowledge is poorer for the questions that are never asked.

The same concentration extends into global governance itself. A recent analysis of 230,000 papers cited by intergovernmental organizations found that fewer than 5 percent of authors account for 30 percent of everything cited by bodies such as the IPCC, the WHO, and the World Bank.

The trap is not sustained by malice. It is sustained by individually rational responses that lock in an outcome no one would choose.

How AI deepens the trap

When AI writing and research tools began to spread, the expectation was that they would narrow this divide. The logic seemed straightforward: if AI lowers the cost of producing a polished manuscript by drafting, synthesizing literature, coding, and providing statistical support, then researchers who previously faced the highest costs should gain the most.

I believe the opposite is more likely, and the more capable AI becomes, the more so. A manuscript is partly a signal. When journal editors and reviewers face far more submissions than they can evaluate deeply, the polish of the writing, the completeness of the literature review, the sophistication of the methods all serve as evidence of the quality of the researcher behind them. AI weakens that signal. Its reach now extends well beyond polish, into analysis, literature, and code, and a manuscript reveals even less. When one signal loses its value, evaluators lean more heavily on the remaining signals, including institutional affiliation, co-authorship networks, citation histories, and prior publication records. These are precisely the signals the recognition trap distributes unequally.

The tools themselves may compound the problem. Forty-four percent of authors in thirty leading economics journals received their PhD from just ten institutions, nine of them in the United States. AI systems trained on this literature plausibly absorb its conventions as their baseline for quality. As these tools enter editorial workflows, research written in a different tradition may face a headwind that no one designed and no one can easily see.

None of this denies AI’s real democratizing effects. Translation, coding assistance, writing support, and access to knowledge once locked behind institutional walls matter greatly for researchers working under linguistic and resource constraints. But these tools help most where the problem is producing work that meets the standard. The recognition gap sits elsewhere: among researchers whose work already meets it. Better tools improve what they produce; they do not change the weight it is given. AI may narrow the production gap while widening the recognition gap.

The foundation at home

The account so far runs in one direction: international recognition unlocks domestic influence. Adnan Khan of the London School of Economics, reading an earlier version of this argument, pointed out that the causality runs the other way as well, and perhaps more powerfully.

Where national scholarship is taken seriously at home — embedded in the institutions that shape decisions, generating its own questions and its own evidence — research acquires a substance and standing that international recognition can then amplify. Where that foundation is absent, recognition rests on individual trajectories with little institutional depth behind them. Some of the deepest failures sit on this domestic side: in much of the world, economic knowledge is simply not taken seriously in policy, with a few strong exceptions such as central banks. International visibility can sometimes help open doors at home, but it cannot compensate for an audience that is not listening.

This matters all the more in today’s fragmenting world order. Fragmentation cuts both ways. It raises the cost of being knowledge-peripheral, because in a transactional world, bargaining power determines outcomes and analytical capacity is a source of it: those who define the evidence define the terms under which economies are assessed, debt is priced, and resources are contracted. But fragmentation also erodes the monopoly of any single recognition hierarchy, as governments navigate among competing frameworks and partners. Which effect dominates depends on what exists at home. A country with real analytical capacity can turn a multipolar world to its advantage. A country without it faces the same asymmetry under every partner.

What breaking the trap requires

A trap held shut by two locks does not open through individual effort. Every actor in the system is responding rationally to everyone else: researchers direct their energy where returns exist, editors rely on the signals available to them, policy institutions draw on the knowledge already validated. No one can move alone. That is why the answer has to be coordination, a mechanism that shifts several margins at once.

Three margins matter most. The first is entry into recognition networks: joint research between scholars in developing-country institutions and those already embedded in leading academic circles changes how work is received — in the same study, collaborations between US-based and developing-country authors reach top journals at 11.8 percent, against 1.4 percent for developing-country authors alone. The second is accumulation: recognition compounds through repeated participation, which is why doctoral pipelines and sustained research partnerships matter more than one-off projects. The third, most often missing, is demand from within: central banks, ministries, and regional institutions articulating the questions that matter to them and working with researchers to answer them. At this year’s Global Empowerment Meeting, several participants pressed a principle that Maroof Syed of CERP put most plainly: solutions must be tailored to problems, not problems to the solutions on offer. That discipline is only possible when those who live with the problems are the ones defining them.

I want to be careful about scale. No single partnership breaks a system-level equilibrium. What a partnership can do is demonstrate that the model works and join a coalition of efforts, from the African School of Economics to new regional doctoral programs, that together can move it. This is the work we have begun between Mohammed VI Polytechnic University in Morocco and the Harvard Center for International Development: joint research on questions defined from African realities, a talent pipeline running in both directions, and a forum connecting researchers to the policy institutions that need them.

Whose knowledge governs

Writing in this series, Gomez Agou has argued that economic sovereignty rests on three foundations, and that the most underestimated is knowledge — the capacity to govern with domestic expertise. The recognition trap is the international face of that same problem. And there has never been a better moment to break it. In a world that is aging fast, the scarcest resource of the coming decades will be youth, and the largest share of it is African. A generation of researchers better trained and better connected than any before it is entering the field, with questions the discipline needs and evidence it has never had. What that generation still lacks is the architecture that turns individual talent into recognized knowledge — and that architecture can be built. Reimagining international development, the theme of this year’s Global Empowerment Meeting, begins with reimagining whose knowledge counts in it.