Iati Scoping Paper - Chapter 4 - Existing Reporting Mechanisms For Donors

Existing reporting mechanisms for donors DAC reporting 1. Donors report aggregate statistics to the DAC database once a year. 1 In addition, they report project details to the CRS, which should be submitted quarterly, but, in practice are reported annually by most. In future, an improved system (CRS ++) will allow the DAC’s aggregate statistics to be built up from detailed project information, and so guarantee consistency, increase accuracy and comprehensiveness and reduce transactions costs. Most donors are now reporting in CRS++ format 2. 2. For most donors, DAC -reporting is a time consuming and labour-intensive process, usually coordinated by a central statistics unit that often has to collate data from multiple agencies and code these manually in line with DAC standards. 3. Despite this effort, the results are far from perfect. The data are published with a considerable time lag, insufficient detail, and partial compliance by donors, and there are persistent issues over the quality and completeness of data. However, these shortcomings should not be seen as a criticism of the DAC, or of individual DAC reporters. Instead, it reflects the relatively low priority given to statistical reporting by some donors combined with A Malawi case study revealed the primary the inherent inadequacies of their reporting means of data collection on future aid systems. allocations to be a spreadsheet-based return that the Ministry of Finance sends out to donors on a monthly basis. Line ministries in Malawi, including education and health, also request data directly from donors on relevant projects to assist their national planning. These parallel requests for information result in discrepancies, with both education and health ministries reporting that the information they were given by donors did not match the information on expected disbursements from the Ministry of Finance, which were generally lower.

1 Preliminary figures for the previous year are published in April, final figures a

re reported October, and the data are released in December

2 In 2007, 15 donors reported in CRS++ format. 12 were DAC members, representing 71% of all DAC bilateral ODA

Other reporting processes 4. AIMs rely on a manual process, asking individual donor country offices for information. This represents a significant transaction cost overhead 3 for both partner country governments, which have to do a lot of chasing and manual entry, and donor country offices, which have to compile the information for AIMs and/or manually enter it in International Development Markup themselves, as well as report centrally. This often results Language (IDML) is a data format led by in inconsistencies between what is reported to AIMs from the Development Gateway for the country offices and from donor HQ to DAC. automated exchange of development 5. The publication of information through other information providers tends to vary: some donors (such as DFID) provide AIDA with an automatic data feed based on IDML (see box), whilst others provide a spreadsheet which is mapped to IDML. In some instances, the AIDA team collects the information from websites using screen scraping technology. 4

project information to populate AIDA. It is an example of a standard data format designed to be machine readable and reusable for different applications. There are other standards, such as SDMX (for the exchange of statistical information), that might be relevant to IATI.

6. Reporting to FTS is also a manual process. A spreadsheet is produced by each donor and sent to OCHA on a monthly basis, and this is manually entered into the database. EC-based donors report to ECHO (EC Humanitarian Office), which then transfers the data to FTS. Reporting to donor websites is usually a separate process and varies from donor to donor. 7. In addition to these reporting requests, donors face a plethora of additional requests for information, from regular organisational and parliamentary reporting, to a wide range of ad hoc requests from NGOs and internationally agencies. 5

Figure 1: an illustration of the multiple donor reporting channels, nearly all of which are manual

3 Development Gateway estimates that manual collection, data validation, and input of data in one country can take up to 2300 days effort from partner (this assumes that government is fully responsible for data entry - the clerical data entry process itself is only a fraction of the total transaction costs) 4 5

country

Governments and up to 225 collectively from donors per year

Screen scraping is a technique in which a computer programme extracts data from the display output of another programme One representative at the DAC working party of statistics highlighted they had 200 requests for data in 2007

Donor Systems 8. All donors have their own internal financial and management information systems, which contain financial and transaction details on aid flows. Most have a separate DAC reporting database, in some cases fed directly from their own systems, in other cases entirely stand-alone. In many cases, the basic data required by the DAC (title, descriptions, country, sectors, dates, commitments, disbursements) already exist within the central financial systems. It is the lack of consistent formats and definitions, the reliance need for data from multiple agencies for complete ODA coverage, the specialist nature of DAC policy markers, and the emphasis on quality control that contribute to the high transaction costs. 9. As previously noted, these donor systems are not publicly available. However, some donors do provide access to certain systems to partner country governments (e.g. World Bank Client Connection, UNDP/UNFPA/UNOPS ATLAS). This presents challenges for governments that need to learn how to use multiple systems, with different definitions.