Our Programmes



Sign up to our newsletter.

Follow ODI

Greening Aid? Understanding the environmental impact of development assistance

Time (GMT +01) 11:00 12:15


J. Timmons Roberts - Chancellor Professor of Sociology and Acting Director, Environmental Science and Policy Program, College of William and Mary, USA.

Bradley C. Parks - Research Fellow, Institute for the Theory and Practice of International Relations, College of William and Mary; Associate Director, Department of Policy and International Relations, Millennium Challenge Corporation.


Camilla Toulmin - Director, International Institute for Environment and Development (IIED)

Seán Doolan - Environment adviser, Africa Division, Department for International Development (DFID)


Neil Bird - Research Fellow, Forests, Environment and Climate Change Programme, ODI

After a brief introduction by chair Neil Bird, one of the authors, Michael Tierney, introduced the book and its related work. The book has come out of a ground-breaking interdisciplinary project known as Project-Level Aid (PLAID) whose objective is to create a web-accessible database on development finance. The final product hopes to include information on all projects committed by bilateral and multilateral aid donors since 1970.  To-date, this database covers 21 major bilateral donors, 40+ multilateral donors, 428,663 projects and a total of $2.3 trillion. Using PLAID ‘1.0’, which includes all projects until 2000, the book offers a satellite-view of environmental aid allocation.

The PLAID effort began in 2003 at the College of William and Mary, USA, and was a response to a strong, early realization that there were many significant gaps in data on aid allocation. Michael Tierney argued that this topic is critically important as it is therefore not clear how much real progress has been made in transferring resources to developing countries in order to mitigate, prevent, or remediate damage to the environment.

Traditionally, aid funding is organized according to OECD Creditor Reporting System (CRS) sector codes, but this coding cannot be used to understand the environmental impact of development projects. Instead, the authors developed their own coding system, which is based on actual project descriptions and does not assume that sectors are necessarily homogeneous. The projects were double-coded into three primary categories:

  1. Environmental Strictly Defined (ESD)Projects (such as access to clean water, biodiversity, CO2 reduction, soil conservation)
  2. Dirty Strictly Defined (DSD) Projects (such as air and road transportation, chemicals, logging, mining, dams)
  3. Neutral (N) Projects (such as banking/finance, disaster relief, education, health, trade).

The 'Environmental Strictly Defined' category was then sub-divided into 'brown' or 'green' projects, where brown projects are those focused on local environmental issues (clean water, sewage, urban environmental issues, erosion control) and green projects are those focused on addressing regional and global public goods (such as climate change, biodiversity, energy conservation).

Timmons Roberts continued the presentation by addressing Question 1 (above): Has aid been greened, and if so, by how much?

  1. Since the 1970s, ‘dirty aid’ has dropped from 55% of aid to ~30%, but this is mainly a function of an overall increase in aid; in real or absolute terms, 'dirty aid' has stayed at around $30 billion/year.
  2. ‘Environmentally-neutral’ aid has sky-rocketed from $15 billion in 1980 to $50 billion, and now constitutes the majority of foreign aid.
  3. Bilateral environmental aid increased by 370% over the 1980s and 1990s.  Multilateral environmental aid also increased by 140%. However, environmental aid remains a small fraction of total aid.
  4. At the end of 20th century, environmental aid leveled off just below $10 billion (approximately 10% of all aid).
  5. For multilaterals, 'dirty aid' receives four times the level of funding compared to environmental aid, and two times for bilaterals.
  6. Both bilaterals and multilaterals began the 1980s with about 10-12 times as much 'dirty aid' as 'environmental aid', but moved in the late 1980s to a much lower ratio. Generally, bilateral agencies greened earlier and stayed greener.
  7. Most multilateral environmental aid has been going to 'brown' (i.e. local public good) projects. Most of the funding has gone to water projects.

Timmons Roberts examined the stated promises at the Rio summit, where the 'Agenda 21' document was designed to break the impasse between developed and developing countries. It called for a significant increase in ‘new and additional’ ODA for global and local environmental problems. Looking at the four main objectives which came out of Rio (water, biodiversity, land and climate), only funding for water came close to the expected amount prescribed. All other categories received only a negligible increase in funding. See Table 2.1 from Greening Aid below:

1960s environmental aid by sector: comparison of agenda 21 prescriptions needed and with actual dose delivered, four sectors.Dose prescribed $b/yrDose received $b/yrPercentage of dose receivedWater6.15.692%Land18.20.352%Climate change200.844%Biodiversity1.750.1257%Total46.056.91515%

Timmons Roberts then addressed Question 2 (above): Which donor governments spend the most on foreign assistance for the environment and why? The top three environmental donors are Denmark, Norway and Germany, respectively. There are many contesting theories regarding why one donor country is greener than another. Some factors that may explain the donor’s commitment to environmental projects include national wealth, so-called ‘post-materialist values’, domestic environmental policy preferences, a strong concentration of so-called ‘green and greedy coalitions’ (such as environmental NGOs), and the strength of a dirty industry lobbying sector.

The PLAID model helps explain why there has been a drop in ‘dirty’ aid rather than a rise in ‘environmental aid’. The wealthier and more post-materialist countries invest less in dirty projects (although not necessarily more in environmental projects).  Moreover, countries with stronger coalitions of the ‘green and greedy’ and higher rates of environmental compliance spend less on dirty aid and more on green aid.

Brad Parks continued the presentation by addressing Question 4 (above): Which countries receive the most environmental aid and why?The top ten environmental aid recipients from 1990-1999 (in order) are: China, India, Brazil, Mexico, Indonesia, Philippines, Egypt, Argentina, Turkey, and finally the Less Developed Countries (LDCs). The authors sought to understand why these countries received the most environmental aid. This was tested across many factors, such as the economic relationship with donor country, the global environmental significance of the recipient country, the colonial legacy, and the severity of local pollution in recipient country. Based on their statistical findings, it appears that:

  1. Countries of global environmental significance receive more green aid from bilateral and multilateral donors
  2. Physical proximity to donor (as a possible proxy for regional environmental significance) is a good predictor of brown aid, but not green aid
  3. Local environmental damage is not a strong predictor, but there are significant measurement problems with this criterion.
  4. Donors appear to screen for recipient credibility (i.e. effective governments, strong environmental policies and institutions) more extensively at the “gatekeeping” stage than the “amount” stage of the allocation process
  5. Bilateral donors favor recipient countries with higher rates of environmental treaty ratification when providing ‘green’ aid
  6. Bilateral trading partners are favored (across all sectors)
  7. Colonial ties matter (across all sectors)
  8. Bilateral donors target poorer countries more effectively than multilateral donors
  9. Bilateral and multilateral donors favor more populous countries 

The presentation concluded with a mention of some limitations of the findings:

  1. the model examines countries as whole units, and does not assess regional variations
  2. the model conflates donor grants and loans
  3. the model does not include any environmental ‘mainstreaming’ into traditional aid projects, so the findings may be underestimating the positive role of aid on the environment

It is important to invest in independent coding and evaluation. For example, the UK Department for International Development (DIFD) reports that during the 1990s, the UK spent 25% of its aid on environmental issues, while the coding in the PLAID model reports only 10%.

The presentation was followed by a commentary on the book, given by Camilla Toulmin (Director of IIED) and Seán Doolan (Environmental Advisor for DFID’s Africa Division). The commentators made the following remarks:

  1. the book includes no references to the work that has been going on in the UK on the greening of aid. This is partly a reflection of the gap between academics and policy-focused practitioners, and the gap in communication between research in the US and the UK
  2. The book is based on an amazing database, but given its size, the quality of reporting needs to be very good. Camilla Toulmin raised an issue of agricultural aid projects all being coded as ‘dirty’, whereas in reality some may be quite environmentally beneficial.
  3. The lack of inclusion of projects which have ‘mainstreamed’ environmental benefits is disappointing. Camilla noted that if mainstreaming is increasing, it should not matter if ‘pure’ environmental projects are decreasing.
  4. For many countries, development aid will become a less important source of finance compared to private financial flows, so there is a need to look at the role of businesses and other financial conduits for making an impact on the environment.
  5. It is critical to look at the role of national political economies and governance structures. Country-level dynamics is critical.

The authors responded to the commentary by saying they regretfully agreed there is a lack of inclusion of UK research in the book. They also commented on the quality of project coding, remarking that if an agricultural project mentioned environmental benefits (such as soil conservation), then it was coded as a ‘green’ project. The authors also remarked that for many countries, aid will continue to be an important source of funding for development and the environment.

The coding database will be made publicly available in the near future.


The recently published Greening Aid analyzes trends in development assistance during the pre- and post-Rio Earth Summit period. The authors have compiled one of the largest datasets of foreign aid ever assembled. By evaluating the likely environmental impacts of more than 400,000 development projects by more than 50 donors to over 170 recipient countries between 1970 and 2001, they seek to answer three central research questions:

  • Which donor governments spend the most on foreign assistance for the environment and why?
  • Why do some donor governments delegate the allocation and implementation of environmental aid to multilateral agencies when they could simply allocate it themselves?
  • And why do some recipient countries receive more environmental aid than others?

Greening Aid explains major trends and shifts over the past two decades, ranks donors according to their performance, and offers case studies that compare and contrast donors and types of environmental aid.  This meeting gave the authors their first opportunity in the UK to discuss their analysis.

Purchase Greening Aid now on Amazon.co.uk