This a very clear and persuasively-written paper that marks a significant advance in the Bank's thinking on public sector management. Very importantly it provides concrete steps in a number of directions, of which these in particular stand out for me: 1. Design and implementation of programs for results. The Approach proposes some critical innovations including (i) diagnosis of functional problems, binding institutional constraints and options- the methodology, provision of tools and on-going support to teams in collaboration with HD look a very promising package (ii) flexibility, adaptability and consideration of the "reform space" including the demand-side in implementation of PSM programs. This latter point speaks somewhat to important political economy considerations that other commentators are raising, and warrants more explanation and emphasis. 2. Explicit and systematic learning. The major program of monitoring, analysis and research to learn how public institutions perform and what works in PSM reform is central to improving results, and has the potential to fundamentally change how this work is done. This work is essential, and facilitating others to contribute is a great idea. The provision of a menu of robust indicators will be very helpful to country/bank task teams, and will support learning from Bank projects – the clarification that these will be optional and can be tailored is very welcome. 3. A holistic view and coordinated approach to PSM. The Approach tackles the critical issue of needing to understand both the upstream and the downstream of PSM – one or other alone provides an incomplete view of the incentives impacting performance, th results chain, how functions perform and resources are actually managed, and what is required to bring about sustainable change. The specific proposals to encourage multi-sectoral team working seek to address a binding constraint for doing this. 4. Setting targets and tracking progress. The involvement of country partners, as well as other development agencies, in developing common indicators would be very positive. A collaborative development process is critical for legitimacy and for collaboration in application, as the paper explains, though it would be important to note that this also means it is likely to be a major task – in effort and elapsed time. There are also significant resourcing implications for data collection – collaborating with others will share these costs, but the high coverage achieved for PEFA has undoubtedly been motivated in part by fiduciary requirements for budget support as the paper notes. A few queries/suggestions: • Standards. Is it necessary or desirable to propose the definition of "standards" against which assessment will be made of "compliance"? "Compliance with standards" may generate a negative reaction, or give the wrong impression about how this would be undertaken and for what purpose. Would anything be lost by instead referring to a common set of performance measures? More substantively, the major challenge in the development of PEFA was less about agreeing on the indicators and more about how the framework would be applied and for what purpose- it was important to balance the desire for monitoring information with domestically-driven processes of change. Deciding specifically on these will be part of the collaborative development process but signaling upfront the intention and tone would be important. • Change management. It would be helpful to more explicitly bring out the need for understanding and adopting effective change strategies and management– in the piloting of new approaches to implementation, and in the learning program, in particular. This, together with the thinking on how to empower stakeholders to create reform space and bring about change, is something on which WBI could contribute. • Consultations. It could be useful to include in the document the collaboration process built into its preparation, involving partners inside and outside the Bank. Getting partners on board to the Approach while it is being developed (and allowing them to contribute) would enhance the likelihood of support, resources and collaboration being forthcoming once it is launched.