Let me try today to answer the questions by Osa (#3 and 4) and John (#5).
Osa asks how the government ensures that external funding are registered as 'on budget /on treasury'. Let me first emphasize that this is not a subject the PEFA program directly addresses, so it is my personal opinions expressed here rather than an institutional position. In my opinion it is largely for the government itself to ensure that this happens. Line agencies invariably like to receive external funds outside the centralized budget allocation and control systems (i.e. funds not counted as part of overall sector allocations) and donors like to have direct interaction with line agencies, rather than having to go through the ministry of finance, as they may be able to demonstrate direct contact with (and implicitly, influence on) sector decision makers and program managers. The central finance agencies may gain control of such arrangements if they control the negotiations and final approval of agreements with external funding agencies and if they control the opening of project bank accounts. So the government will need to ask itself if the ministry of finance has got the legal powers to exercise this control and if so, whether the ministry of finance effectively enforces those provisions.
On your other question, 'government driven' or 'donor driven' aid processes are two sides of the same coin. The government can only drive the aid process if it is willing to assign the required resources (staff time) to take the lead and if it is willing to forfeit certain aid allocations if the relevant donor(s) is unwilling to fit in with the government's agenda. If the government fails on those two counts, the aid processes easily become donor driven. Donors, on the other hand, may help the process by providing sufficient time at different stages of the process for the government to consider work done and take its decisions on priorities and allocations on that basis, often with political considerations as an important input. This is one reason why PEFA assessments do not include recommendations or action plans. Action plans (including revision of existing ones) are clearly expected to emerge from a PEFA based PFM assessment, but if they are already included in the assessment report, they have a tendency to be based only on technical considerations, and to address all areas where weaknesses are found. That approach to action plans often leads to too many recommendations/actions with insufficient consideration of the polital realities of reform - with the effect that the reform do not get implemented. It is important that the government has enough time to digest the description of current strengths and weaknesses and decide for itself what reform priorities it wants to set - based on what is seen by the government as the most urgent improvement needs, and what vested interests the government is willing and capable to confront at any one time.
And then to John's question (#5). Users' impression of PFM performance from the sectors they are familiar with, is a fact of life. One reason why this may affect the acceptability of the indicator ratings is that sector variance for a particular indicator rarely is described in the PEFA assessment report. This is clearly a weakness, because many of the PEFA indicators require that system performance is assessed at sector level (in a sample of sectors) before these performance assessments are combined into an overall indicator rating across government ministries and agencies. In some cases, PFM system are fragmented and each major component needs to be assessment separately e.g. (for PI-18) there may be different payroll systems in operation for education, health, security services and the remainder of public servants respectively. In other cases, user opinions on general PFM performance may be influenced by the sector's input mix e.g. the education sector performance becomes associated with how personnel is managed, whereas the transport sector may be associated with how large public works contracts are managed - even if the same type of system performs uniformly across government entities. The PEFA program is working on guidelines for how to strengthen the sampling of sectors as a basis for arriving at overall rating for the relevant indicators, and how to present the results as evidence for the ratings in the report, without significantly expanding the work of an assessment and the length of the report.