The centrality of collective action problems in governance for development: New evidence


This page in:

Evidence is piling up on the need to revisit the standard ‘supply’ versus ‘demand’ concept of how to improve governance for development. This is pointing to an exciting set of new priorities for reform in sub-Saharan Africa.

From the side of intervention experience, Margaux Hall’s excellent contribution on this site, Peeling the Mango, invites a much fuller and broader discussion. From a research angle, the recently published report from the Africa Power and Politics Programme on Governance as a Collective Action Problem proposes a broader way of framing this experience. I will say more on the research presently. What’s the issue with the mangoes?

‘Peeling the Mango’ deals with social accountability efforts in Sierra Leone. Margaux Hall’s blog post points to a serious mismatch between what local facilitators find effective in working towards improved health service delivery and the theory that in principle governs their intervention.

The programme theory suggests an adversarial relationship between the ‘supply’ and ‘demand’ sides of service delivery. Citizens are pitted against state agents, encouraged to monitor service providers and demand their rights. The practice, in contrast, takes into account the penalties that may be incurred by villagers who make formal complaints. It also recognises that service providers face their own constraints, often struggling to meet heavy demands with severely limited resources. Facilitators find that it helps not to use adversarial methods and that it pays to invest in understanding the specific problems affecting provision. This leads them to focus on brokering mutual commitments among providers and users, which sometimes results in better provision.

Margaux offers this as an illustration of the importance of understanding socio-cultural dynamics when working to strengthen accountability relationships in local settings. It is that, but it has further implications.

To begin with, this is not an isolated example or one that reflects the particularities of Sierra Leone. An equivalent mismatch between experience and programme theory is cropping up in local accountability interventions in several parts of Africa, with recently documented examples from Malawi and Tanzania. In those settings, too, significant local improvements have been achieved with non-adversarial facilitation approaches, sometimes without any explicit recognition of the methods actually being used, particularly when reporting to funders.

Secondly, the Sierra Leone case illustrates a broader point about the current status of the ‘good fit’ agenda, around which Brian Levy launched this blogsite. As Brian argued, it is time to go beyond the mantra that every country is unique. And, as Merilee Grindle pointed out last year, advocates of good fit need to start providing development practitioners with robust middle-range guidelines on what to do differently, as distinct from studying every situation afresh before doing anything. One of the implications of this line of thinking is that, if current programme theories are being refuted in practice, we need to be putting something in their place. In other words, this is a moment for new concepts and directions.

In the field of social accountability, we do have some new concepts. The idea of ‘working across the demand and supply dichotomy’ is now much discussed, thanks to the build-up of documented experience on ‘blurring the boundaries’ between state and non-state action in improving service delivery. Research in the Citizenship and Future States centres at IDS Sussex have contributed importantly to this awareness of problems and potentialities.

Now, however, further research is questioning the usefulness of merely bridging ‘supply’ and ‘demand’. The above-mentioned synthesis report of the Africa Power and Politics Programme (APPP) argues that the supply/demand metaphor is itself problematic, reflecting the persisting influence of an unrealistic principal-agent perspective on governance reform in Africa.

As the report presentation and summary put it, governance challenges in Africa are not fundamentally about one set of people getting another set of people to behave better. They are about both sets of people finding ways to act collectively in their own best interests. In other words, it needs to be recognised that actors at all levels typically face collective action problems in acting in ways that favour development results.

This applies to the government leaders tacitly treated as developmentally oriented ‘principals’ in the best-known types of ‘supply side’ public sector reforms. It applies equally to the citizen groups treated as principals in the many intervention logics that rest on the power of information and eliciting demand for good governance. The APPP research shows that thinking about governance reform as a collective action problem is not just helpful in making sense of what works in the provision of local public goods. It also sheds light on the big questions about why some types of national political regime promote economic transformation and others do not.

The Sierra Leone, Malawi and Tanzania examples suggest that, in some important respects, development practice is running ahead of the relevant theory. That may be welcomed as sign of high-quality programme implementation. But it is a less than ideal state of affairs in that it wastes the potential for systematic learning.

For example, we are told that the government of Sierra Leone is evaluating the impact of the intervention by means of a randomised control trial. That is good, but only on condition that the exercise is serious about specifying what is being evaluated. Is it the stimulation of citizen ‘demands’ and rights-claims, in adversarial mode? Or is it what the intervention teams consider the better approach: facilitating constructive relationships among providers and users based on a detailed appreciation of barriers to behaviour change in specific local settings? Unless this question is answered, we shall not have learned much about what works. In this important sense, theory matters.



David Booth

Research Fellow, Overseas Development Institute (ODI)

Margaux H
October 31, 2012

Thank you for this thoughtful post and for continuing the discussion of practical and theoretical challenges and opportunities in social accountability practice. I too view the supply/demand metaphor as inappropriate in many circumstances. As just one example, a nurse does not act solely on behalf of the government. She wears different "hats" as she balances relationships at the professional, personal, and community levels. She herself may need to demand better governance if, say, she isn't on government payroll or has a chronic shortage of essential drugs. Effective social accountability programs will grasp her particular circumstances and the potential tensions within her agency on behalf of the state. On the one hand this suggests that we must invest in understanding context at the micro-level; yet I do not feel that we need to study every situation completely afresh. Rather, we can engage in smarter research -- we can aim to learn why interventions succeed or fail, and use those lessons to design more effectual social accountability programs. In other words, we can invest in understanding processes of change, rather than designing towards predetermined end states. In our Sierra Leone evaluation, the government, World Bank, and NGO partners are doing just this. A randomized-controlled trial is evaluating empirical improvements in the quantity and quality of health care delivery, as well as evolving clinic-community relationships – whether in an adversarial or constructive posture. And supplemental qualitative research is probing the dynamics of change, whatever form that change may take. We do not privilege one form of engagement in the evaluation. We look forward to sharing the results and joining hands with practitioners and scholars like you invested in comprehending pathways of change.

Albert van Zyl l Open Budgets Blog
November 23, 2012

Interesting post, thanks. Most of the examples that you cite seem to be from work at local government level though. One wonders whether the reason why local officials are to collaborate is that they think CSOs can help them influence national government? My guess is that just like not all advocacy challenges are principal-agent problems, not all of them are collective action problems either. Look at the Treatment Action Campaign in South Africa. Or the the NCDHR's campaign 789 in India.

December 10, 2012

At a more micro-level, I'm aware of work done with an interesting tool, the Local Governance Barometer, created and piloted by Pact, Idasa, and SNV - a good write-up of its use in Madagascar is on pages 154-171 of the book Social Accountability in Africa and should be available at….

What's interesting about it is that the tool emphasis the local conversation about problems rather than the utility to outsiders of a rigorous assessment of the situation. It clearly involves significant roles for outsiders in mobilization, so it's not as great a recipe of a self-generating process, but it is an interesting example. Some background on the tool's development, from an earlier point in time, can be found at

It explicitly borrows from Pact's organizational capacity development process in supporting the facilitation of a broad conversation about governance, though linked to a generalized model that frames the discussion. Like a capacity assessment, the tool is not primarily about measurement, but is a process intervention to generate discussion, reveal degrees of consensus or disagreement across lines of citizenry and government, and come up with concrete actions to move forward. I'm only aware of it being used in certain communities, but to good effect. Hope you find it interesting, and thanks for the commentary!