What Factors Predict a Donor’s Aid Transparency?

The concept of transparency appeals to aid scholars, practitioners, and policymakers alike as a mechanism for improving aid targeting and aid effectiveness. Its positive effects have been well-documented: increased coordination and specialization by donors, smarter budget planning, and greater accountability from recipient governments. However, while some donors have been early adopters of the open data practices by regularly releasing large stores of valuable data, many have lagged behind, and some have effectively opted out of the aid reporting regime altogether. If almost everyone seems to agree that aid transparency is a good thing, why are some donors more transparent than others?

This question brought me to the 2012 Aid Transparency Index, compiled by Publish What You Fund (PWYF), an organization that “monitors the transparency of aid organisations in order to track progress, encourage further transparency and hold organisations to account.” Their annual index of 72 donor organizations measures compliance with the International Aid Transparency Initiative (IATI), the leading international standard for publishing foreign aid data signed by over 30 states.

Using Publish What You Fund’s 2012 Aid Transparency Index, I tried to predict aid transparency at the donor level. My initial model includes: GDP per capita, school life expectancy (the age at which students stop attending school), regime type, the 2012 Corruption Perceptions Index (CPI) compiled by Transparency International (with higher scores indicating lower perceived corruption), and the 2012 Web Index published by the World Wide Web Foundation, which assesses 61 countries’ Web usage, impact, and utility from 0-100 (with higher numbers indicating greater Web impact, connectivity, and infrastructure).

The following scatterplots suggest positive correlations between a donor’s Aid Transparency Index score and GDP, CPI, and Web Index, respectively.

 
 



The results of an ordinary least squares (OLS) regression suggest that the model explains 62% of the variation in the dependent variable, and we can therefore assume the independent variables explain some variation in aid transparency (see OLS regression results here). Interestingly, only the Corruption Perceptions Index (CPI) is statistically significant. When one holds all other variables constant, the model suggests a 7.83 increase in the PWYF variable for every one unit increase in CPI. This finding suggests that donor countries plagued by high levels of corruption (where the government's commitment to transparency is presumably weaker) are generally less willing to disclose detailed information about their overseas aid activities.

Unfortunately, the model I developed suffers from a lack of robustness due to a small sample size (only 17 donors were covered by all indices). A larger sample and more elaborate experimental design would more accurately predict a donor country’s commitment to aid transparency. As more donors support and implement IATI standards, future iterations of the aid transparency index will include more observations (both more donors and more years of coverage), thereby providing a stronger basis for broadly applicable conclusions. Furthermore, other variables likely affect aid transparency, including geographic location, bureaucratic structure, presence of violence, and public sector ethics. Therefore, further research is needed to identify the factors that determine a donor country’s level of aid transparency.

Elsa Voytas (’13) is an AidData Senior Research Assistant at the College of William & Mary.

Tags: citizen voicestransparency