Thanks for sharing this paper. I wanted to point out a few things that I feel should be incorporated into the text to ensure accuracy and completeness.
1) It would be helpful to note that, in many countries, AMP has already been integrated with IATI data (see our report here for more details: http://www.developmentgateway.org/2016/06/28/use-iati-final-report/) and that there is a free and open source tool specifically for integrating AMP (and other AIMS) with IATI: http://www.developmentgateway.org/2016/03/31/launch-iati-import-tool/ - in fact, the Uganda AMP team was just recently trained on the use of this tool. This tool was also repurposed for the recent integration of UAMP with DMFAS in Uganda: http://www.developmentgateway.org/2017/02/23/uganda-dfmas/.
2) I am curious as to the measures used to determine that IATI data are more disaggregated than AMP (in particular as the disaggregation measures used in the table above this comment show comparability between IATI and AMP data)
3) Comparing the number of projects within IATI and AMP may not be a useful exercise, as IATI often contains multiple representations of the same project (e.g. NGOs may report on a project they are implementing for World Bank, who may have received funding from 5 bilaterals, resulting in 7 records for one project).
4) In general, it seems that a more in-depth analysis of data quality (see for example our blog on this topic here: http://www.developmentgateway.org/2015/03/17/iati-and-country-systems-data-evaluation-methodology/) would be welcomed. It is important to ensure that using IATI adds value (we firmly believe that it can/does), and in which DPs/cases it does so, before calling for a wholesale shift to: "The starting point should be to implement IATI-AMP automatic data import; then join AMP with other local systems such as budgets." It is important that decisions like these also take into account i) the needs of the government (as the primary users of AIMS), and ii) a detailed analysis of data quality and tradeoffs.