Cross-posted from the IATI website

Bangladesh has long been a champion of IATI – as Vice Chair of the IATI Steering Committee, as a contributor of USD 100,000 to the Secretariat and as an endorser of the Initiative for over five years. It is not surprising therefore that Bangladesh is keen to use IATI data to enhance its Aid Information Management System (AIMS).

As the Government of Bangladesh (GoB) uses a locally developed AIMS and owns the source code, they are able to flexibly and cheaply make whatever changes and improvements to the system they want. With the support of UNDP, they have contracted a great local IT company (Technovista) to develop a way to easily import IATI data to the AIMS.

Getting good data is a challenge for all AIMS

AIMS are important tools to capture information about development cooperation at a country level and are vital platforms for coordinating and overseeing activities funded and implemented by a large number of different actors. To do this they need basic data on the projects all actors are planning and implementing.

Manual data entry takes time and effort, and data quality suffers as a result. The increasing availability of IATI data is an opportunity to lower the burden of data entry and get higher quality data available for stakeholders in aid recipient countries to analyse.

Challenges in using IATI data

There is now a large amount of data published in IATI. Given this, questions have been raised about the very limited use of IATI data in country systems. The lack of use has been a source of frustration for organisations that have invested considerable resources in IATI and in their own data publication. Our work in Bangladesh is identifying some of the difficulties faced by enthusiastic partner countries:

Complexity in the aid system: Country level systems need to be able to handle the same project being reported by multiple organisations from different perspectives. Multiple reporting is encouraged in IATI as publishers provide data independently of each-other, and because a much wider range of organisations (including implementers) can publish. This complexity is inherent in the aid system – where money is frequently pooled or passed around between organisations – and there are significant benefits to having multiple reporters. However, in IATI the burden of coordination (removing double counting) is currently shifted downstream to the end users. If more publishers correctly used the traceability components of the IATI Standard this situation would improve. For the foreseeable future, AIMS that use IATI data will need to rely on humans – particularly staff in donor country offices – to interpret the data.

Differing donor approaches to providing IATI data: Data quality is not a precise term and varies wildly depending on the ‘for what/who’. At country level, key quality features are whether details such as the implementing agency is named, and translations given in the national language. Some donors publish data with aggregated financials or without grouping data into recognisable projects, or publish data too infrequently to be useful.

We will not be implementing any donor-specific ‘hacks’ to deal with these difficulties. Instead, we will contact donors and inform them of these problems, encouraging them to address them. There should be a strong business case here for donors to address these issues, as it should save a large amount of staff time from manually re-keying data that could otherwise be imported. The forthcoming publication of Publish What You Fund’s 2016 Aid Transparency Index should also help to focus minds on data quality.

Due to these challenges, IATI data doesn’t yet flow to recipient countries as smoothly as might be imagined. However, this will not prevent us from using the wealth of good data that is now available. Our solution is to build an interface to allow humans to make sensible decisions when importing IATI data, explaining how activities fit together, and at what level, to create meaningful projects, adding in missing data and clarifying that status of the data with in-country representatives as they go.

Where we will get to by May

We have a very tight (and ambitious!) timescale for development, which must be complete within the next three months. By then, we aim to have developed an interface which can import data from any organisation publishing IATI data. Donors will continue to be responsible for reporting their projects to the Bangladeshi AIMS and it will be up to them to decide whether their IATI data is an accurate reflection of their activities or whether they prefer to use other data sources. We will provide tools to help them make these decisions.

We are also conscious of the opportunity to create some useful public goods out of this process. The software Technovista are developing is open source and openly licensed: it will continue to be published on Github as we progress. The methodology, documentation, and development process is being recorded on a microsite. We are also using this platform to collect the issues we find in using donors’ data. This is useful for us to refer back to and point donors towards, but we hope it will also be a helpful resource for others working on importing IATI data to country systems in the future.

Moving from data collection to aid effectiveness

2016 can be the year that we demonstrate the power of IATI to reduce the burden of data collection, improve data quality and, ultimately, use the data to improve aid effectiveness. We’re working hard to make that possible at the country level, and we will be happy to be held accountable for delivering on our commitments at the IATI Members’ Assembly meeting in June.

We can only work with the data that’s available – and some donors still need to catch up with the rest on the quality of data they are publishing. But with a wealth of good-quality data now out there, it really is time to start using the data that is available – and rewarding those donors that have made progress with a lighter workload.