Editor’s Note: The following piece appeared in Transparency and Public Procurement, a thematic supplement to the 2011 Annual Statistical Report on United Nations Procurement.
Introduction
There is growing consensus that aid transparency must be improved in order to increase aid effectiveness. According to Moon and Williamson, aid transparency can be defined as “the comprehensive availability and accessibility of aid flow information in a timely, systematic and comparable manner that allows public participation in government accountability.”
This includes not just how much money is given, but how that money is spent, and as such is of great importance when studying the public acquisition of goods and services in countries which receive Official Development Assistance (ODA). In this paper, we look at elements of transparency that are needed to improve aid coordination and accountability, from the donor to the procurement officer.
The Accra Agenda for Action, a document summarizing the deliberations of the 3rd High Level Forum on Aid Effectiveness held in 2008, called on all donors to disclose aid information in a timely manner. The International Aid Transparency Initiative (IATI) was launched at the same forum. IATI brings together donors, recipients, aid experts and non-governmental organizations to create a common and universally agreed method of sharing aid information between all stakeholders.
Importance of transparency in aid effectiveness
Complete transparency means that everyone can see how much aid is being given by each donor, to whom, for what projects and when.
Over the last ten years, the number of new aid projects has skyrocketed, and their average size has shrunk drastically. This fragmentation of ODA makes it even harder for aid agencies to coordinate their activities and duplication and waste could be growing.
Transparency is also valuable in combating corruption. In one beneficiary country in Africa a public expenditure tracking survey found that only 20% of donor-funding for education programmes was actually reaching schools. As a result of an information campaign making transparent what each school was supposed to receive from the ministry, the funds flow increased to more than 80%.
Greater transparency affects all major stakeholders in development assistance programs:
- Donor country taxpayers can understand how their taxes are being used, and thus become more engaged in and supportive of aid.
- Donor country governments can evaluate their aid programmes more effectively.
- Recipient country citizens can hold their governments to account over any discrepancies between aid received and aid spent on the public procurement of goods and services.
- Recipient country governments can plan their budgets and their procurement needs better. This is especially true for aid dependent countries, where ODA forms a large part of their budget.
Improving transparency will require additional investment and organizational changes by donors. These costs are mainly administrative, adapting IT and reporting systems to global standards and also staff time spent on training and the reporting of aid activities. But the savings on automating provision of aid data alone would more than offset the costs of investing in better transparency systems. The case against transparency thus seems to be fundamentally political – a reluctance to automatically release information that could potentially be damaging to an organization.
The need for a transparency index
A transparency index serves as a quantitative measure that is comparable across countries or agencies. A benchmarking system assesses a donor against what others are actually doing in practice rather than against an abstract notion of ‘good’ behaviour. A secondary purpose is to enable research to document the importance of making progress on the transparency agenda.
Almost everyone pays lip-service to the importance of transparency but without specific indicators it is hard to hold donors and implementing agencies accountable for putting their commitments into practice. A transparency index fills this gap by generating a dialogue on which donors are putting a transparency agenda on aid into place and how aggressively they are moving to implement such an agenda.
The transparency index data and methodology
To create a useful transparency index, we need to focus on data that agencies provide to publicly available, comparable databases.
There are three defining characteristics of a strong transparency index. First, it should only use data that donor agencies proactively put in common databases, such as IATI, so that it can be accessed and compared with other donors. The notion is that transparency implies data that is readily available, useful to others and comparable across donors so that it can be a tool for greater coordination and accountability. Information buried in an agency’s annual report or web-site does not meet the requirement of crossdonor comparability.
The second important characteristic of an index is that it differentiates between complete and partial reporting. In this way, we are able to make judgments on the overall quality and comprehensiveness of aid information and make more nuanced judgments on the degree of transparency.
Third, it should explicitly compare donors to each other, thereby creating a ‘best in class’ measure and a base year measure of transparency.
The data for the transparency index can be taken from two main sources: the Development Assistance Committee’s Creditor Reporting System’s database; and AidData, a data source for aid activities launched in March 2010 as an independent organization not affiliated with any donor group.
The existing datasets are poor in a number of regards. First, the datasets are incomplete, especially with regard to coverage of the most important donors from emerging economies and of private aid. Also, variables like disbursements cannot be accurately matched with commitments, so it is hard to know if projects are actually implemented. Second, the datasets are not timely, with up to two years’ delay. Third, there are areas which could be important for transparency but where all donors do poorly. Two such examples at the time of writing in 2011 are geo-coding and beneficiary feedback.
The transparency index we designed is one component of a broader effort at assessing aid quality introduced by Birdsall and Kharas that is continually being updated. In its initial construction, our transparency index was an equally weighted average of six indicators, each of which is directly actionable. They were chosen to reflect what can best be thought of as a ‘culture of transparency’ that we believe is linked to aid effectiveness. The indicators are defined as follows:
- Whether the donor is a member of the international aid transparency initiative (IATI) – an initiative to agree on common standards and reporting to facilitate sharing of aid information.
- Proportion of projects for which the project title, its long description and its short description are filled out in the AidData database.
- Average character count of the project long description in AidData. Although lengthy descriptions are not needed, in general the more detail on the aims of the project is better.
- Percent of projects reporting the aid delivery channel. It is important to know whether a project is to be implemented by the government, an NGO, a multilateral or another agency.
- Completeness of project level commitment data. In our analysis, we have found a discrepancy between the reported amount of aid at the aggregate level (what the donor claims to have donated in total to a given beneficiary in a given period) and the sum of aid at the disaggregated, project level (reported disbursements for all individual projects taken together in the same period). The missing or unaccounted aid is not transparent by definition.
- Share of net ODA that donors give to recipients with a good monitoring and evaluation framework. This indicator rewards donors that support countries with good monitoring and evaluation frameworks.
Useful findings revealed by a Transparency Index
Based on these indicators, it is possible not only to rank and compare donors but also to spot overall trends. For example, in our initial compilation in 2011 there did not appear to be any strong, systematic difference between multilateral agencies and bilateral donors on transparency. There was a very low correlation between the size of donors and the transparency of their activities. Several large donors were also poor performers on transparency. By 2012, we found that some donors had made a forceful effort and overall global transparency was being improved at a faster pace than any other dimension of aid quality.
Donors who are members of the IATI score higher on other dimensions of transparency as well, suggesting that IATI members are broadly committed to transparency. Indeed, 13 of the top 15 most transparent donors in 2011 were also members of the IATI. View the full ranking and other related research on www.aiddata.org.
Conclusions
It is not straightforward to measure transparency, partly because norms and standards are still not universally accepted and partly because transparency is an elusive and shifting concept that resists an easy definition.
This paper proposes an index of transparency based on six indicators. We hope that the benchmarking provided by these indicators will help to ‘move the needle’ in the transparency of agency activities. The downside of the chosen indicators is that we are restricted to available information, which is currently limited in scope.
Greater transparency in aid would help reduce overlap, waste and the lack of coordination between donors. It would also help beneficiary countries plan their public procurement needs and hold their procurement officers, and other holders of the purse strings, to account. Lack of transparency also leads to a lack of opportunities to learn what really works in aid, thus inhibiting rigorous research on aid effectiveness. Because aid is increasingly fragmented, norms and formalized systems of transparency are becoming more important. Informal knowledge sharing among a few large players is no longer a viable alternative, as ever more players need to know what others are doing.
