A new paper (pdf) from UNDP is very old wine in a new bottle. It purports to estimate the “cost” of halving poverty in 15 of the poorest African countries. It concludes that aid in the amount of about 13-25% of GDP could cause poverty to be halved there over the next nine short years.
Sounds good: Pour in more money and poor people will become richer. Unfortunately, it just doesn’t work that way. In the thicket of equations, you’ll find two assumptions for which there is zero evidence: All aid becomes investment, and all investment becomes income. No serious economist believes that---not even close---which makes this exercise totally irrelevant to aid policy.
While controversial, there is certainly some evidence of a positive relationship between aid and investment (see the work of Hansen and Tarp) but nowhere near a one-to-one relationship, especially in the poorest countries. And the evidence for a strong relationship between investment and income in the poorest countries of Africa is totally lacking (see the work of Shanta Devarajan and William Easterly).
It’s hard to stomach sentences like this one, from the paper:
“Although Zambia has received an amount of $86.50 (the largest amount among the 15 countries) for foreign aid for the past decade, it requires more aid to meet the goal of halving poverty by 2015”.
Bill Easterly has calculated that if all aid given to Zambia had become income, as the UNDP paper assumes, Zambia today would be as rich as the United States (see The Elusive Quest for Growth: Economists' Adventures and Misadventures in the Tropics). Numbers that retain such ridiculous assumptions should claim none of your attention.
The only way to move forward is to innovate massively, work like crazy to improve aid quality, and find better ways to support recipient countries in improving their own policies. We do not need another “MDG costing” study---a genre of research that should die with this paper.