Research Systems, and How Can Partnerships Strengthen Them?
October, 2025
By Francesco Obino, Global Development Network (GDN)
When we talk about research systems, we are often trying to capture the bigger picture that shapes the daily realities we know so well as researchers, administrators, or funders. The term refers to the material, institutional, and political conditions that make research possible—or constrain it. Understanding these systems is the first step to changing them.
Defining Research Systems
There is no single, neutral definition of a research system.
- The Doing Research global initiative at the Global Development Network (GDN) defines it as the interplay—at the national level—between the production, diffusion, and use of research.1 This focus reflects GDN’s mission: supporting local research capacity as a development strategy in its own right.
- UNESCO, instead, emphasizes the constellation of institutions, policies, and actors that enable knowledge creation and use (2015).
- The OECD refers to national innovation systems that connect research with education, enterprise, and society.
What these definitions share is an ambition to shift the level of analysis: from the success or failure of individual projects or researches, to the conditions that enable them (and knowledge more broadly) to flourish.
Why Research Systems Are Difficult to Strengthen
Researchers are products of the institutions that train, employ, and fund them—and those institutions are, in turn, shaped by broader policy choices, incentive structures, and specific historical legacies. Changing these conditions for the better is never straightforward. Success at one level often creates imbalances at another: funding elite researchers can widen inequalities; long-standing institutional partnerships may crowd out new players; and setting trends in research topics will leave others neglected.
In today’s “project economy,” the assumption that many successful research projects will automatically strengthen a research system is fundamentally misguided. The opposite is likely true, and the imperative to “do no harm” applies to research collaboration and funding just as it does to any other field. (Obino, Salomon and Zanfini, 2023).
Knowing these conditions to start with, is not an obvious task. Each system has its own history, trajectory, forces at play and constraints, and grasping these nuances requires deliberate effort. In Niger, trade unions have been key to reforming research careers (Marou Sama, 2016). In Myanmar, more than 170 universities coexist with very limited research output (GDN and CESD, 2021). In Benin, 93% of social science researchers are unaware that the country has a national research funding agency (GDN and ACED, 2025).
Partnerships as Entry Points to System Thinking
Collaborations are an ideal entry point to engage with research systems. No partnership can claim to be equitable—or even effective—without understanding the local systems in which each partner operates. Knowing how a collaborator’s environment shapes their incentives, autonomy, prospects, ideas, and understanding the constraints they face compared to one’s own, is a condition for meaningful partnerships (see P2). It can be transformational to inform the design of collaborative projects, the terms of collaboration, and the prospects for equitable implementation.
Rethinking Capacity as the Matter of Partnerships
Capacity is not a binary quality—strong versus weak—but a layered and situated dimension of careers, institutions and systems. Capacity strengthening means uncovering and engaging with these dynamics. Thinking about capacity—especially in the terms of those whose capacity is under discussion—means focusing on how they perceive their own system, not just on what they are assumed to be able or unable to do. It can be counterintuitive, too (see P4).
A department might lack gender balance because another donor is already funding all the women in the university; a think tank’s visibility may hinge more on its ability to build and retain a team of research assistants than on its CEO; a university’s credibility can depend entirely—or not at all—on a single event; an obscure publication can change a career; transdisciplinary research may stall simply because no one feels competent to review it; and a shortage of quality proposals for a funding call may reflect the quality of the call itself, or the reputation of the funder amongst researchers.
This shift in the way capacity is discussed within partnerships is essential, and it requires humility. Each collaborator must be willing to discuss their own capacity, and what strengthens or weakens it within their system. When done well, such discussions become deeply engaging and enabling, while also illuminating research systems (see P3 “Know your Partners and Collaborators” and P4 “Contribute to Strengthen Institutions”).
Rethinking Impact and Accountability
The pressure to demonstrate short-term “impact” has made project design more technocratic and often less transformative. Projects that genuinely aim to strengthen systems should be accountable both to their own stated results and to the broader public good of improving research ecosystems—or at least, of not harming them. This demands impact literacy not only from researchers but also from donors, who must recognize that systemic change may take longer and be less visible, yet is ultimately more enduring. Practically, this means talking less about impact, particularly at the project level, and more about the position of projects and their achievements in the research systems that they happen in (see P4 “Consider your Impact on Research Systems”).
A Practical Way Forward? Debating and writing about research system within partnerships
The more we know about research systems, the easier it becomes to strengthen them. One simple rule of thumb could make a major difference: if every research collaboration devoted even 5% of its resources to understanding and documenting the systems in which it operates, the global “project economy” would already take a decisive step toward building stronger systems.
Such discussions—about funding mechanisms, incentives, and governance—are deeply engaging, often transformative, and too rarely written down. Embedding them in project design, whether as an inception step, a dedicated work package or a cross-cutting focus, helps leverage collaboration for cumulative learning.
Caring about research systems is another way of caring about the conditions that make research possible, equitable, and impactful. Partnerships are the optimal space to mainstream this reflection—helping us design better projects, manage expectations, allocate resources fairly, and strengthen the environments we all depend on. The effectiveness of partnership to do this depends largely on the partners’ capacity to self-critically situate themselves, and the collaboration, in the research systems they operate in.
If every collaboration began by studying the system(s) it inhabits, as a natural step in understanding each other, we might finally turn a patchwork of successful projects and global research partnerships into an engine of systemic change.
- “Social science research system: The set of institutions, practices, structures and rules that enable the production, diffusion and uptake of social science research. This document uses the terms ‘research system’ and ‘social science research system’ interchangeably.” (GDN, 2017:7)
References
Further Resources
UKCDR (2022) Four Approaches to Supporting Equitable Research Partnerships
The Global Development Network (GDN) is a public international organisation that supports high quality, policy-oriented, social science research in Low- and Middle- Income Countries (LMICs). It does so by promoting research that is contextualised and locally driven, leading to more informed policies, increased policy ownership, better-informed implementation, and more sustainable and inclusive development choices. Since 2014, GDN runs the Doing Research initiative, and coordinates a community of researchers committed to document their respective research systems, across more than 30 countries.
Learn more about GDN: