Editing ODA: What to Omit and Add in the Definition of Aid

My earlier posts on the definition of Official Development Assistance (here, here) dwell on the questions of when and how to count loans as aid. That focus makes sense because it was a controversy over loans that drove the donors of the Development Assistance Committee to ask the staff to review whether there is a “need to modernize the ODA concept.” In this post, I’ll set loans aside and make recommendations on what kinds of activities to drop from or add to ODA. For details, see sections 1 and 2 of my CGD policy paper.

What to Omit

In 2005, the UK advocacy group ActionAid made a splash with its Real Aid report. It alleged that “two thirds of donor money is ‘phantom’ aid that it is not genuinely available for poverty reduction in developing countries.” For example, debt relief for poor countries was a good thing, but signing documents in Paris to put old debts to rest doesn’t in itself feed or house the poor. And when donors hire their own nationals to train census takers or advise on power plant construction, recipient nations don’t see a penny of that money either.

As long-time idea-peddler, I admire Real Aid for its impact. Within a couple of years, the DAC expanded its aid statistics system to flag the sorts of spending excoriated by ActionAid. This then let the DAC zero in on aid money that actually crosses the border into a developing country, which they call Country Programmable Aid; it is a subset of ODA. To this day, the Real Aid report frames much of the debate over ODA definition, to the extent that some kinds of aid have become conventionally controversial. But I think that, in its drama, the report overreaches.

In my paper, I favor keeping some of the controversial categories and dropping others. I say:

•  Keep technical assistance. Yes, sometimes advisors give useless advice, but sometimes their work pays dividends, when they implant valuable knowledge in the minds of powerful, public-minded people Just like any other kind of aid, sometimes technical assistance works and sometimes it doesn’t. Nor is TA distinctive in being a kind of export of donor-country goods or services. All aid ultimately supports exports, being delivered in kind or in hard currency.
•  Keep administrative costs. It’s easy to lambast administrative overhead as bloat, unworthy of inclusion in the ODA tallies. In fact, I subtracted it from the first edition of the Commitment to Development Index. But Mark McGillivray showed me the error of my ways. Administrative functions such as monitoring, evaluation, and auditing are part and parcel of aid delivery. They are not excess.
•  Probably keep counting the subsidies in admitting foreign students to domestic university at low tuition rates. Surely if many of the students return after exposure to a new world of ideas, this can be transformational. I always think of Manmohan Singh, who studied economics at Cambridge in the ‘50s and presided over reforms as India’s finance minister in the ‘90s, which added trillions of dollars to cumulative national income. Whether the 1950s equivalent of ODA financed his studies, the mere possibility makes the point.
•  But don’t count the cost of hosting refugees in donor countries. It is laudable to house and feed people have just arrived in wealthy nations with little more than the clothes on their backs. But for statistics, a sensible line has to be drawn between foreign and domestic aid. And since most refugees, like my wife, will stay, aid to them crosses that line.
•  And exclude most debt relief. See recommendation 4 of my previous post.
•  And don’t count programs to educate citizens about global poverty. Maybe promoting “development awareness” in wealthy nations stimulates private overseas charity as well as public support for official aid. But the activity itself is not aid.

What to Add

•  Count development loans with near-market interest rates, such as from the World Bank’s flagship lending program, the International Bank for Reconstruction and Development. Recommendation 3 in my previous post explains.
•  Count guarantees and payments for performance such as COD aid based on provisioning. Traditional ODA consists mainly in grants and loans. Grantors take no financial risk; lenders take risk but fear it. One thread in the ODA redefinition discussion asks how to embrace financial instruments that themselves embrace risk. For example, some development assistance agencies issue guarantees to cover private lenders against default on loans for projects developing countries. CGD’s Nancy Birdsall, William Savedoff, and Rita Perakis advocate cash on delivery aid, which is like a guarantee except that it pays out on success rather than failure. Agencies making such contingent commitment need to set aside funds to prepare for possible payouts, called provisions. And they need to report these transactions on their financial statements. I favor counting as ODA the credits to provisioning accounts.
•  Count private overseas charity that is attributable to tax incentives. The Commitment to Development Index has long done this. For example for 2011, the method attributes $11.5 billion of the recorded $23.3 billion in private overseas charity in DAC nations to income tax deductions and credits that reward charity, and to lower taxes in some countries, which leaves more money in private pockets for private charity.
•  Count 100% of contributions to the UN peacekeeping budget, not 6%. The CDI does that too. Peacekeeping looks like aid: it costs money, can promote development, and takes place largely in developing countries. Surely posting 10,000 blue-helmeted troops to Liberia in 2003 to maintain a fragile peace contributed to that country’s development.

My spreadsheet tool lets you see what dropping the “conventionally controversial” categories does to donors’ aid totals. Most of the changes would not be dramatic. Administration accounted for 5.1% of ODA in 2011, technical assistance 4.6%, and the other categories for less. The proposed additions would have similarly incremental impacts.

Nevertheless, the ODA concept needs to evolve. If it does not, it will only fall more out of step with the times, putting its credibility at risk.