I hope that Duncan Green's comments don't discourage other practitioners from reading this report ("Development as a Collective Action Problem: Addressing the real challenges of African governance") – because it is a very good read. It is true that it doesn't compromise on academic rigour, but that is all to the good, given – as the report itself puts it – the propensity of donors to derive upbeat messages from a selective reading of key bits of evidence. While the report's argument is complex and nuanced, it is clearly and carefully developed, and weaves detailed empirical findings drawn from extensive fieldwork into a compelling bigger narrative. It presents the APPP research in the context of wider relevant research findings, with the result that it serves as a very useful survey of current evidence on governance and development. The paper does rehearse now familiar arguments against "best practice" approaches, but it goes on to ask a fundamentally important question: what does the alternative of a more context specific, "good fit" approach look like? Or, to put it another way, which institutional patterns and governance arrangements work well, and which badly, in supporting the provision of public goods that matter for successful development? Inevitably the answers offered are partial, and are more compelling at a local than a national level. The paper offers some actionable findings and ideas about practical next steps: but its more important contribution is to challenge (sometimes quite brutally) the half-baked assumptions on which much current development activity is based, and to offer clear alternative ways of thinking about core development challenges. The paper contains some crucial messages that make it essential reading for researchers and policymakers. To select a few: - We need a more nuanced approach to thinking about neo-patrimonialism. The paper identifies some key factors explaining variations in development performance under neo-patrimonial regimes, including how political competition shapes the way elites use rents. - It is important to focus on economic transformation, not just growth – not least because this is needed to support effective democratic governance; and without such transformation, multi-party competition risks reinforcing damaging "competitive clientelism". - Perhaps most importantly, the paper argues that it is unhelpful to think about development and governance challenges from a principal-agent perspective: as the author memorably puts it: "governance challenges are not fundamentally about one set of people getting another set of people to behave better". Employing a principal-agent framework to analyse governance problems results in simplistic "supply" and "demand" side approaches, and deters policymakers from investigating the real underlying constraints to effective collective action by both public and private actors. This is a really powerful message for policymakers, with some uncomfortable implications for many current demand-side interventions designed to strengthen citizen voice and empowerment, and demands for better government accountability. - The report emphasises the need for stark realism about the huge challenge of getting effective collective action in poor countries without capitalist economies, with fragmented elites, and with multi-party political competition that reinforces short termism. It makes a brave – but tentative – attempt at suggesting some "big picture" solutions, including an element of power-sharing to reduce the negative impact of "winner takes all" political systems. - There are more compelling policy messages about problem solving at a local level, including the identification of some potentially actionable obstacles to collective action. These include policy and institutional incoherence, exacerbated by inconsistent and fragmented donor interventions, and populist initiatives by governments. The report has good, detailed evidence from fieldwork of effective local problem-solving that is locally anchored in prevailing social values and institutions, but which also draws on specialist skills of public agencies: a nuanced version of "working with the grain". - There is some very hard-hitting (and mostly well-deserved) criticism of the actual and potential damage caused by ill-conceived aid interventions: these can add to policy incoherence, promote approaches (for example democratic decentralisation) that worsen problems of public goods provision, and often employ associational models and funding that undermine local collective action. While the paper identifies the primary challenge as being to convince ministers, parliaments and the voting public in the North of its key messages, I agree with Duncan Green that, as currently written, it is unlikely to do that. It would be well worthwhile preparing a version specifically for this audience, and Duncan's suggestion of brainstorming the research findings and recommendations with a wider group of practitioners would almost certainly make for a more compelling final section.