A more vivid PLAID
Weaving a PLAID:
The big room at the Corner House holds a number of William and Mary students categorizing foreign-aid data under the supervision of faculty such as Mike Tierney and Rob Hicks. (Photos by Steve Salpukas)
Support from Gates and Hewlett foundations accelerates version 2.0 of aid-tracking initiative| January 9, 2009
PLAID is moving to the next level.
PLAID—the acronym for Project-Level Aid—has received nearly $2 million in financial support from the Bill and Melinda Gates Foundation and the William and Flora Hewlett Foundation. The funding will allow researchers to increase the size and depth of the project. Faculty involved in the project characterize the advancement as “moving from version 1.0 to 2.0.”
An interdisciplinary program dedicated to offering detailed, accessible information on the flow of international aid, PLAID was formed at William and Mary in 2003 as a project to compile and maintain a web-accessible database on international development finance. It tracks foreign aid between sovereign nations, known as bilateral aid, as well as multilateral aid, which refers to financial assistance provided through multinational organizations such as the World Bank and the United Nations.
The PLAID team is led by three members of the College of William and Mary faculty, Robert Hicks, J. Timmons Roberts and Michael Tierney. The team is interdisciplinary, as Hicks is from the economics department; Roberts is from sociology; and Tierney is from government. A parallel group of lead investigators are faculty at Brigham Young University, and includes Daniel Nielson, Darren Hawkins and Sven Wilson. The final member of the team is Brad Parks, a research fellow at William and Mary’s Institute for the Theory and Practice of International Relations and an associate director at the Millennium Challenge Corporation.
Looking at gaps in the data
Parks, a 2003 graduate of William and Mary, wrote a senior honors thesis on foreign aid and the environment that sparked the formation of PLAID. His examination of the changing allocation of aid by Western governments revealed gaps in the data being used by development researchers all over the world. PLAID began filling in the missing information and addressing inconsistencies on records of thousands of individual projects.
“PLAID is a unique project for me because it combines the missions of academic research and public service,” Hicks said. “This project will hopefully impact, in a positive way, the lives of the poor all over the world and at the same time help shed light on some of the fundamental questions in development research.”
To construct PLAID 1.0, researchers compiled and re-categorized more than 430,000 individual aid projects from 1970 to 2000. Tierney, associate professor of government and director of the Institute for the Theory and Practice of International Relations, explained that PLAID 2.0 will contain updates, additional data and a deeper level of detail.
“PLAID 2.0 is going to merge a variety of different data efforts,” Tierney said. “First, we’re going to update through 2006 by gathering data on development projects from 2000-2006 for all of the donors we have in PLAID 1.0. Further, we are filling data fields in PLAID 1.0 that are currently not populated.”
Tierney said that among the most important of the version 1.0 fields to be thoroughly completed throughout 2.0 is one holding a “long description” of the project. Existing databases, he explained, often lack a detailed narrative account of what the project is supposed to accomplish.
“So, it’s quite difficult for a researcher to know really what was going on in the project,” he said. “There may be a title and a short description, but a four- or five-sentence description is missing in a lot of these records, so we are finding and then adding long descriptions for PLAID 2.0.”
New categories of information
In addition to filling in the blanks of 1.0, PLAID 2.0 will include new categories of data to reflect changes in the way aid is administered. Traditional foreign-aid packages, Tierney said, were often delivered for a single purpose—to build a road or to recruit teachers for a rural school, etc. PLAID 1.0 was able to accommodate single-purpose grants with a single set of categories, but the international aid world has outgrown that approach.
“In real life, and increasingly, donors give what are called multi-sector projects. USAID will give a project to Tanzania, but it won’t just be to improve their heating and power generating. It will be to improve their heating and power generating and build a school and rope off the rain forest,” Tierney explained. “How do you categorize that? If you’re insisting on providing just one category for the project, you’re in trouble, because you don’t know how to weight it. The new PLAID 2.0 has multiple sector categories. So you have a primary sector category based on the budget and you can categorize each chunk separately. This gives you a much more marbled, much more nuanced view of aid flows than 1.0.”
In addition, the new and improved PLAID will also include more donors, especially emerging bilateral donors from Eastern Europe and the Middle East. “In the last five or ten years, many more sovereign donors have gotten into this game of providing development assistance,” Tierney explained, citing Poland, Lithuania, the Czech Republic, Kuwait and Saudi Arabia as examples. “Now, even Venezuela, with their oil money, and China and India are emerging donors.”
Tierney pointed out that PLAID 2.0 will also include emerging multinational donors. PLAID 1.0 had 15 or 20 multilateral donors—the biggies, such as the United Nations and the World Bank. The upgrade to 2.0 will include about 25 more obscure, but active, multinationals, such as Scandinavia’s Nordic Development Bank and the North American Development Bank.
“Who knew that the North American Development Bank even existed?” Tierney asked. “Well it turns out that it has been in existence since NAFTA was formed. It’s a multilateral bank and it’s designed to provide funding for development projects along the border between the United States and Mexico. Most of their projects are funded in Mexico.”
Most importantly, PLAID 2.0 will be a step toward the open, accessible and searchable database originally envisioned back in 2003 as a valuable tool for donors and recipients and others interested in the flow of bilateral and multilateral aid. Tierney expects a beta version of PLAID 2.0 to be released in March, 2009 to a group of target users: academic researchers, personnel representing both donor and recipient governments, journalists and people from non-governmental organizations (NGOs).
“We want these five groups to bang on the database, to beat it up, to use it in a way that they would do their own work,” Tierney said. Recipients of the beta version will get together with the PLAID staff in September 2009 at a conference at the University of Oxford to share their experiences with using the 2.0 database and the PLAID interface and to discuss how to make both more powerful and user-friendly.
“We want to build PLAID 2.0 so that it is useful to a wide variety of users. PLAID 1.0, which was funded by the National Science Foundation, was basically built for the academic community. But Gates and Hewlett were very insistent that PLAID be useful to lots of different people,” Tierney said. “So, donors might use it for aid coordination, to find out what other donors are doing in a particular country. Recipient governments might use it to do planning: ‘Aha! Look! I can see now that Germany is giving out more aid for infrastructure. We’ve got this infrastructure project. We need to approach Germany.’”
Merge and purge
The construction of PLAID 2.0 is a matter of merging and purging, with a lot of checking in the middle. Rob Hicks is the member of the team who has taken charge of what Tierney calls “the grand merge.” The merge includes the PLAID 1.0 database and incorporates a database from the collaborators at Brigham Young, where the multiple-sector concept was developed. It also includes an updated version of the database from the Organization for Economic Cooperation and Development, a Paris-based organization that has been the traditional collector of and repository for foreign-aid data. PLAID 2.0 also will contain information about aid projects from a number of smaller databases.
The melding of databases has generated additional research opportunities. Roberts led a group of PLAID students this summer categorizing 115,000 projects from the updated OECD database on aid. The students looked for projects that were designed to help countries cope with climate change or to reduce their emissions.
“There is a major swing in development finance internationally to try to ‘climate-proof’ development projects, by preparing recipient communities for changes coming their way.” Roberts said. “‘Climate’ is on the lips of the donors: The World Bank just set up two new ‘climate investment funds’ with donations from ten donor nations worth $6.1 billion.”
Still, there has been no systematically maintained set of information of just how much development funding was flowing to climate work. The William and Mary team, led by undergraduates Tommy Jones and Kara Starr, found that funding for energy efficiency and hydroelectric projects dwarfed funding for "adaptation" to climate change. "These findings entirely support the claims many are making—that there is barely any funding going to helping the poor nations of the world prepare for and recover from the natural disasters which scientific models predict are taking an upswing with warmer and more extreme weather events like droughts, hurricanes, and rising sea levels," said Roberts. "Now researchers from many agencies and academic institutions are requesting to search PLAID data for these kinds of patterns, and we aim to help them do so as soon as we can."
“We are in the process of merging it all together,” Tierney said. “You can merge a lot of it automatically, because there are unique features in the project line items, but for hundreds of thousands of projects, the computer cannot match them. Those hundreds of thousands of projects get flagged for human matching. So if you walk into the next room right now, you will find student research assistants in there, doing line-by-line matching.”
He explained that the computer merge will generate two projects that look very similar. A student will examine them and apply a set of rules to see if they are different projects or computerized versions of the same project, one of which should be purged from the database.
“Basically, the students are doing quality control on the merge,” Tierney said. “We have very, very tight quality control, because people are concerned that we’re going to get duplicates or that we’re just going to miss stuff when we do the merge. This is another great reason to have the Oxford conference. People who work on development projects will be in the room. They’ll be able to point out that their water project in Guatemala isn’t in the database, or it’s in there twice—or it’s classified incorrectly. Getting that feedback from people in the donor community and the recipient community will be very, very helpful in vetting the quality of the data.”