As many of you know, we here at AidData have made it a priority to pursue information on emerging donors and South-South Cooperation partners. We've seen how Arab donors can have a huge impact in a single country, or how Brazil is leveraging comparative advantages from its own experience to assist other developing countries. But there's one unique case that stands out from all the rest, allowing us to add a new donor and a new recipient in one blow.

This post was written by Eva Baker ('12), Mike Tierney and Michael Weissberger.

UPDATE: Sunday, August 29th:  For a different look at the issue of historical and forward-looking aid flows to Pakistan, see the excellent post by Molly Kinder and Wren Elhai from the Center for Global Development entitled "Pakistan Aid Facts."

That post is a great compliment to this one because it does a number of different things that we don't do here, such as:

1. While our historical data on U.S. assistance to Pakistan ends in 2008, Kinder and Elhai show U.S. aid projections up through FY 2011 budget request.  One very cool thing about their data (from the U.S. State Department) is that even before the floods the U.S. was planning a massive shift the sectoral allocation of aid to Pakistan away from health and education toward infrastructure projects.  One of the points of our original post was that Arab donors focused on infrastructure, whereas Western donors have not.  It appears that may not be true going forward.

2. While our data on aid to Pakistan comes directly or indirectly from donors, Kinder and Elhai also use data from the recipient (the State Bank of Pakistan), which presumably has a pretty good idea of who gave how much money to Pakistan (though this might not count some off-budget development assistance).

3.  Point #2 illustrates a shortcoming of AidData data for certain donors.  Some donors, like Saudi Arabia, do report project level data from some of their government agencies (the Saudi Fund for Development).  But they don't publish project level data (or in some cases even aggregate data) from Finance Ministries or line agencies that also allocate development assistance.  This point is addressed by Debra Shushan and Chris Marcoux here.  This may explain the huge differences between estimates of Saudi assistance to Pakistan over the past decade.  It turns out the Saudi Fund for Development has only given about $30 million over the past decade, but when all grants from all Saudi government sources are included, Pakistan reports receiving about $137 million per year between FY 2004-2009.

4. Kinder and Elhai have cooler pictures.

Speaking of cool pictures, this graphic might cause critics of Saudi stinginess to change their tune.  When calculated on a per capita basis, the Saudis are among the most generous donors to Pakistan's flood relief -- second only to...Kuwait.

Original Post from August 24...

The humanitarian crisis caused by the recent floods in Pakistan highlights both how important international assistance is in any response and recovery effort, but it also illustrates how much more efficient such a response could be. Growing estimates of the objective need (4 million homeless, many more at risk of malnutrition, waterborne disease and exposure) contrast with a shortfall of humanitarian assistance. Estimates of the amount of aid required for the relief effort have skyrocketed to around $1 billion over the coming months, which does not count the reconstruction costs.

These current appeals should be understood in the context of historical aid activities in Pakistan. It makes sense to know which donors have been providing development assistance in the recent past and which donors have “boots on the ground” in Pakistan. Much of the immediate assistance being provided by external donors is actually aid that has been re-programmed to assist in this emergency or it is in-kind assistance being provided by foreign military forces (primarily U.S.) located in the region. Therefore, it makes sense to know who the traditional donors are and whether they are operating in Pakistan. It also makes sense to know where these development projects are located at the sub-national level, since this would dramatically enhance the ability of Pakistan's government to coordinate the donor response and would enable donor organizations on the ground to efficiently re-program their own assets. Unfortunately, to date donor organizations do not report the location of their aid activities in any systematic manner.

AidData has the most complete historical picture of international assistance to Pakistan. These data include aid from traditional OECD DAC bilateral donors, multilateral development banks, multilateral grant-making agencies, and non-DAC bilateral donors (such as Saudi Arabia, Poland, Brazil, Kuwait, etc…).

Figure A illustrates the total volume of development finance dollars flowing into Pakistan between 1999 ($2 billion) and 2008 ($6 billion). These flows include all aid from official bilateral aid agencies and from multilateral aid agencies that are tracked by AidData. The massive increase in aid during 2003 is partly the result of debt rescheduling, where most of the traditional donors either forgave some portion of Pakistani official debt, or they restructured the debt to provide better terms. The apparent drop in 2009 should be ignored since not all donors have reported their aid flows for 2009 and 2010.


There has been much debate about the role of Arab donors in Pakistan since the floods hit a few weeks ago. Figure 1 provides a comparative view of development assistance committed by Arab donors and from the United States, the largest bilateral donor to Pakistan over this time period.

In 1998, after Pakistan tested a nuclear weapon, the U.S. suspended foreign assistance to Pakistan so that only a small amount of humanitarian aid was flowing. However, after September 11, 2001, the U.S. needed Pakistan's cooperation in order to conduct military operations in Afghanistan. Consequently, the U.S. dramatically increased foreign aid (and military assistance) to Pakistan. (The military assistance is not shown on any of these graphs). U.S. foreign aid to Pakistan increases dramatically from around $100 million in 2000 to over $1.3 billion in 2003. The amount of aid increases in the later part of 2001 and stays very high through 2008.

AidData tracks aid from 3 different “Arab bilateral donors,” including: Saudi Arabia, Kuwait, and United Arab Emirates. AidData also tracks aid from 4 multilateral agencies that are financed primarily by Arab governments, including: Islamic Development Bank, OPEC Fund for International Development, Arab Bank Economic Development in Africa, and the Arab Fund for Economic and Social Development. The last two of these multilateral agencies allocate money only to Africa and to other Arab countries. Since Pakistan is neither an Arab nor African country, we should not expect, nor do we observe any aid from these organizations. However, the other 5 Arab donors have all given some development assistance to Pakistan over the past 10 years. The ISDB has consistently given the most, while Kuwait, Saudi Arabia, the UAE, and OPEC have given relatively small amounts of aid in some years, and no aid in other years.


While U.S. aid has stayed relatively high over the past 8 years, ranging from $200 million to $1.3 billion per year, Arab aid to Pakistan shows a clear downward trend. By 2008, only Saudi Arabia and the ISDB were making new aid commitments to Pakistan. (2009 and 2010 should be ignored due to slow reporting by some donors.)


In addition to variation in the amount of aid, we also observe dramatic differences in the type of development assistance given to Pakistan by the U.S. as compared to Arab donors. Figure 3 shows the different sectors where the U.S. allocated aid to Pakistan over the past 8 years. About 60 percent of the aid has come in the form of debt relief or direct budget support. This is the most fungible type of assistance -- cash. The next largest sector is Social/Health/Education spending, followed by Food and Humanitarian assistance, which flowed in large quantities after the October 2005 earthquake in Pakistan.

Arab donors allocate their assistance overwhelmingly to infrastructure projects. Whereas the U.S. only provides 4% of its aid for infrastructure, the Arab government donors provide 100% of their aid in the infrastructure sector. The multilateral Arab donors are only slightly more diversified, but again the largest single sector for both OPEC and the ISDB is Infrastructure (as illustrated in Figure 4 and 5 below).



We put ISDB on a separate graph (Figure 5) because it gives so much more money than the other four and because its portfolio is the most diverse. Still, infrastructure aid dwarfs the other sectors of the ISDB portfolio. These aid flows to Pakistan are consistent with previous studies, which find that Arab donors tend to fund infrastructure projects, such as road building, airport construction, hydro-electric facilities, telecommunications, and mineral extraction. While other donors invest a larger proportion of their aid into humanitarian assistance and rapid response activities, Arab donors have traditionally been more active in the reconstruction phase of emergency response. Hence, they are often criticized for being “slow to respond” when there is a Tsunami in Indonesia (2004), an earthquake in Pakistan (2005), an earthquake in Haiti (2010) or the current massive flood in Pakistan (2010). The news coverage is very consistent, if not always fair or accurate, that oil rich Arab countries do not help quickly enough and they are not as generous as they could be. These criticisms are especially strident when the emergencies occur in predominantly Muslim countries.

However, the increased aid from Arab donors one or two years after such emergencies might simply reflect the fact that these donors have experience doing infrastructure work in their normal development projects – hence, their comparative advantage in the relief and reconstruction effort focuses on reconstruction phase.

The 2010 flood in Pakistan has provoked considerable criticism among both Western commentators and Pakistani voices about the slow response from oil rich middle Eastern donors. However, after two weeks the Saudi royal family, which is difficult to distinguish from official Saudi government aid, did pledge to raise $74 million, which is the second largest pledge behind that of the United States at $156 million.

The timeliest data on aid commitments and disbursements to relieve flooded areas of Pakistan (and other global humanitarian aid flows) is published by the UN OCHA Financial Tracking Service. The folks at the UN working on that project report official legal commitments, but they also report pledges that are "uncommitted." That system is updated daily and accessible in this spreadsheet to reflect new promises and actual dollars that are allocated.

This post was written by Brad Parks, a Research Associate at the Institute for the Theory and Practice of International Relations at William and Mary and Principal Investigator on the AidData project.

A growing number of scholars and policy wonks argue that foreign assistance is characterized by a "broken feedback loop": donor agencies and development banks are not, strictly speaking, accountable to their beneficiaries and they generally have weak incentives to disseminate information about the efficacy of their programs. This distinguishes foreign assistance from other types of government spending, such as domestic social welfare programs.

One potential remedy is to make it easier for project beneficiaries and local stakeholders in developing countries to share real-time performance information with Western donor agencies. If EuropeAid's local contractor uses shoddy materials to build a school in Tanzania, there is no reason (in principle) why a parent should not be able to go to an internet cafe and register a complaint, or better yet, use his or her cell phone to submit a picture of the school's leaky roof. Before you write this off as the utopian ideal of starry-eyed humanitarians and pointy-headed academics, check out CGD's recent interview with Rakesh Rajani. Rakesh explains in some detail how explosive mobile technology growth in East Africa has substantially increased transparency and accountability.

We are pleased to announce that AidData has launched a new initiative to enable project beneficiaries, NGOs, journalists, and other interested parties to augment donor-provided data with first-hand observations, photos, videos, news reports, and documents. Here's the basic idea: by mashing-up the official data reported by donors with 'observational' data from people on the ground, we would like to leverage the AidData platform and facilitate conversations between donor agencies and their intended beneficiaries about the performance of specific development projects. Our hope is that these new "enhanced project view" pages will provide real-time project performance information and a shared space where donors, NGOs, and other interested parties can share completion reports, impact evaluations, and key lessons learned. This virtual space would provide the foundation for realizing many of the principles articulated by OpenAid.

We are very excited about this new initiative, but we need your help to a) create a useful and engaging enhanced project view design, and b) establish "proof of concept" for crowd-sourcing information about donor-funded development projects. We have developed a "mock-up" based on an MCC-funded project in Nicaragua. We invite readers to send us specific suggestions to improve the layout, content, and functionality of these hypothetical webpages. Also, If you know of NGOs that might be interested in working with AidData to evaluate the real-world prospects for crowd-sourcing aid information, please let us know by sending us an email to info@aiddaa.org.

It remains to be seen whether the "wisdom of crowds" can be used to strengthen accountability in donor-recipient relationships, but we think it's worth a try.

The following is a post by two of our research assistants, Alena Stern (William and Mary '12) and Josh Powell (BYU Public Policy Graduate School '11)

Responding to long-held concerns about uncoordinated donor behavior, the Paris Declaration of 2005 made harmonization one of its five pillars, pledging to work towards “eliminating duplication of efforts and rationalizing donor activities to make them as cost-effective as possible.”

Among the problems of uncoordinated action are elevated transaction costs as recipient governments struggle to comply with variegated donor rules and procedures (Knack and Rahman 2007) and decreased donor specialization as “all donors seem to want to give to all sectors in all countries” (Easterly 2007).

All of the attention given to coordination problems begs the questions: how much do donors coordinate their activities? And is coordination effectively targeting needs within a country – both spatially (aid to villages or provinces) and sectorally (aid for different purposes)? Recent work by the World Bank – AidData partnership (Mapping for Results Initiative) sheds new light on the topic.

Using new geo-referenced aid data, we considered the spatial and sectoral coordination of all of the currently active projects of two donors – The World Bank and the African Development Bank – within Kenya and Mozambique. An AidData query places the World Bank and the AfDB as the first and twenty-third highest active donors to Kenya and Mozambique. Because of their influence, analyzing the funding patterns of these two donors offers a sense of where a large proportion of the funding to Kenya and Mozambique is going. (It also calls attention to the need to mainstream geo-referencing for the other approximately 50 active donors.)

World Bank and AfDB aid flows to Kenya demonstrate a tremendous concentration of aid from both donors in the Mombassa-Nairobi-Lake Victoria corridor. This is not entirely surprising, as these are the main population centers of Kenya, but this corridor is also comparatively much better off than the more arid North and East, which receives virtually no aid from either donor. Neither donor appears to be coordinating efforts nor effectively targeting the neediest areas of the country.

World Bank and AfDB aid flows to Kenya

The situation in Mozambique paints a different picture: AfDB projects are primarily directed to the northern parts of the country whereas World Bank projects are clustered near the capital city, Maputo. As in Kenya, however, there appear to be manifold financing gaps in the poorest parts of Mozambique, especially in the northwest provinces - Niassa and Tete - and the provinces just north of Maputo – Gaza and Inhambane. Geographic coordination in Mozambique is encouraging, yet the areas of greatest need appear to be neglected still.

World Bank and AfDB aid flows to Mozambique

Using the same active World Bank and AfDB projects, we then examined sectoral coordination. In Kenya, we find a high degree of sectoral specialization between the two donors. The AfDB is largely focused on central government budget support, allocating 58% of its spending towards this sector. By contrast, the World Bank’s emphasis has been Transportation (29%) and Agriculture (17%).





While the production sectors, such as transportation, power, and agriculture, receive a great deal of aid from both donors, sectors targeting human capital, such as education, health, and social services, are relatively neglected by both donors.




Sectoral coordination in Mozambique is quite different from Kenya: the Power and Transportation sectors are targeted heavily by both the World Bank and the AfDB. But beyond these two sectors, the World Bank and the AfDB spend on quite different purposes as shown below:




In Kenya, uncoordinated donor specialization may neglect particular sectors and geographic regions as the World Bank and the AfDB seem to target the same geographic area and the same sectors. In Mozambique, the donors specialize in different sectors and different areas. By cross referencing this finding with the distribution of sectorally-specific indicators of need, we could assess whether the different types of aid received by northern and southern Mozambique are tailored to the specific needs of those areas. It is possible that the geographic variation in aid portfolios is matched to the geographic variation in need. However, without more complete analysis, including a wider group of geo-coded donors, it is difficult to draw strong conclusions about the significance of the revealed patterns of sectoral and spatial coordination. Areas and sectors that appear neglected by the World Bank and AfDB might be covered by other donors. We would love the opportunity to test this proposition, but that would require that more donors geo-reference their own projects or provide access to their project documents so that we could geo-reference their projects as we have done for the World Bank and AfDB.

Geo-referencing development projects is one very promising method that could facilitate greater coordination and dialogue between policymakers and development practitioners, as well as researchers, NGOs, recipient governments, and citizens. As geo-referenced data become more common, donors and other interested parties (like us) can use the data to find and publicize areas and sectors that are not receiving an amount of aid that is proportional to need (measured in various objective ways). Such visualizations do not provide clear answers about where aid should be allocated, but they raise important questions with striking clarity and will encourage donors and recipient governments to explain allocation patterns to beneficiaries and taxpayers. We think this will improve the prospect of aid dollars arriving where they are needed most.

We're always working to expand the breadth and depth of AidData's coverage, and today we've added projects from four new donors: UNESCWA (the United Nations Economic and Social Commission for Western Asia), UNESCAP (the United Nations Economic and Social Commission for Asia and the Pacific), The World Bank's Carbon Finance Unit, and what we're calling The World Bank Managed Trust Funds.

• UNESCWA weighs in at 5 projects, all from 2007, which total $115,441 (USD 2000).  The source for these records is the OECD CRS (Creditor Reporting System).
• UNESCAP brings in 6 projects, also all from 2007, which compose $76,325 (USD 2000).  The source for AidData's UNESCAP records is the OECD CRS.
• The World Bank Carbon Finance Unit reports 148 projects from 2000-2010, bringing in $1,487,007,719 (USD 2000).  Our source is the donor website.
• The World Bank Managed Trust Funds is a catch-all for several different World Bank funds.  We will be separating these funds out into their own donor records soon.  Until then, the trust funds bring in an additional 347 projects from 1998-2010, and a total of $2,380,203,343 (USD 2000).  The sources for these records are also the donor websites.

These additional donors represent an additional $3,867,402,828 (USD 2000) of aid in 506 individual projects.  Stay tuned here for more updates as we continue to expand our donor coverage and detail in the coming weeks.

The following was written by two of our William and Mary research assistants, Anne Bernier ('11) and Eva Baker ('12)

While in Morocco this summer, Anne heard a lot about the government’s recent work to improve the status of women. In early 2004, the Moroccan Parliament overhauled the Moudawana, or Moroccan Family Code, giving women greater rights. The new family code removed provisions requiring women to obey their husbands and restricted polygamy. The reforms changed inheritance laws, allowed women freer access to employment and education, permitted Moroccan nationality to pass to children through either men or women, and empowered women to initiate divorce.

As Research Assistants forAidData, we wondered how Morocco’s donors would react to the domestic changes. As Nick Kristof has written, aid to women can have dramatic development effects.Did gender-related aid increase after 2004, perhaps to complement the reforms or as a “reward”? Or, did progress in the domestic legal code cause donors to shift aid away from women’s projects into other areas? For a quick comparison we keyword searched Morocco’s aid for the terms “gender,” “women,” “female,” and “girls” to flag projects with a significant women’s component in addition to those with women as the priority purpose. While the method isn’t perfect, as a first cut, we established a replicable and reliable sample of projects that would allow us to establish variation in allocation over time.

From the data we learned that despite a net increase in total aid to Morocco from 1999-2008, gender- related projects received around 1% of total aid both before and after the 2004 reforms. So, while gender- related aid did increase from $10,698,525 in 1999 to $16,307,016 in 2008 (and the number of projects went from 12 to 71 per year), we don’t see a dramatic change in allocation priorities by Morocco’s donors in either direction. This surprised us—donors had been clamoring for these reforms for years; but when they were passed we observe no obvious response on the part of these donors. This outcome is dissimilar to that noted by Chris Marcoux in the environmental issue area where change in aid is related to change in the domestic institutions of the recipient country.



But while there has been no major shift toward gender-related aid relative to other sectors in Morocco’s aid portfolio, perhaps donors became more targeted in the gender aid they were giving by focusing more directly on women’s rights, rather than traditional development activities thought to benefit women disproportionately. Thus, in our continuing quest to find a donor “response” to the reforms, we instead looked at the composition of projects within gender-related aid:


This shift from education and business to civil society projects is similar to the United Nations Development Fund for Women’s strategy. UNIFEM proposes to politically empower women by increasing political representation and mobilization and improving domestic legislation and its implementation. Domestic and international efforts in Morocco could also be viewed within the larger movement to raise women’s status as a critical component of economic development, as noted by Isobel Coleman in Foreign Affairs.

But these descriptive statistics say little about the actual results for women in Morocco. Other countries and women’s rights groups lauded Morocco for its reforms, but despite instances of progress, change has been slow. In particular, many women’s NGOs despair spotty implementation by conservative judges and the national government. To track the reforms’ results, we used Womanstats to look up the UNDP’s Gender-related Development Index (GDI), which uses literacy, school enrollment, life expectancy, and income levels to calculate the social and economic disparity between men and women. A high GDI shows a higher level of achievement for both men and women and higher gender equality.


Since 2003, Morocco’s GDI has risen steadily, although its rank has bounced around from 100, to 95, to 111. So while the achievement levels of Moroccan women have increased since the Moudawana reforms, Morocco has stayed roughly in the same place relative to the rest of the countries in the world. Although we found that donors reacted to the reforms in 2004, it was in a more nuanced way than we expected, given their sweeping changes. Whether that was appropriate will have to be debated elsewhere.

One of the surprising findings in Greening Aid: Understanding the Environmental Impact of Development Assistance was that, the greater a country’s participation in environmental treaties, the less environmental aid it received (controlling for a variety of other factors). This was surprising because the authors expected that recipient country credibility would influence allocation of environmental aid by donors. The basic idea: To achieve the greatest environmental benefit, donors ought to favor recipients who have a demonstrated commitment to environmental protection. Ratification of environmental treaties is one signal of such commitment. Recipients with little environmental commitment might implement projects poorly (or not at all). One possible explanation for this finding is that donors use environmental aid prospectively rather than retrospectively. In other words, donors allocate environmental aid in order to build capacity for environmental governance in countries where current capacity is low.

This reading of the evidence is supported by an intriguing new paper by Michaël Aklin and Johannes Urpelainen of Columbia University. In “The Domestic Preconditions for International Policy Diffusion: The Case of Environmental Ministries,” Aklin and Urpelainen examine how domestic variables and international factors have influenced the creation and diffusion of environmental ministries in countries across the globe. The authors argue that “developing countries have strong incentives to establish environmental ministries when they (i) are undergoing a democratic transition and (ii) the international salience of environmental problems is high.” Aklin and Urpelainen use AidData to calculate yearly flows of environmental aid, which is taken as a measure of the international salience of environmental problems. There isn’t space here to discuss the full range of empirical work that the authors present; I strongly encourage those who are interested in this field to read the entire paper. One especially interesting finding is that democratization and higher levels of environmental aid (measured as 3-year rolling averages) do make the creation of an environment ministry more likely. If environmental aid is given prospectively, to build capacity for environmental governance in recipient countries, this is exactly what we would expect to find. Democratization enhances overall prospects for capacity building, and environmental aid provides both incentive and resource for developing environmental institutions. Figure 8 shows how environmental aid and democratization interact to make the creation of an environmental ministry more likely.