Waylaid Dialectic

August 13, 2010

Inadequate Arguments Against Aid (hoisted from the old blog)

Filed under: Aid — terence @ 9:05 pm

[Hoisted from old blog and modified slightly]

Over at Crooked Timber John Quiggin deals with a common aid fallacy: ‘we’ve given them billions of dollars in aid and they’re still poor – therefore aid doesn’t work’.

It’s commonly observed that despite receiving over $500 billion in aid in the 50 years since the shift to independence, Africa is still poor and, on the whole getting poorer. The implication is that the aid must have been wasted or stolen, and of course, quite a bit of it has been. But an aggregate number over 50 years isn’t very helpful.Much more useful in thinking about the likely impact of aid is the amount per person per week. With (very roughly) a billion people in Africa and a billion in the developed world, the aid that’s been given amounts to about $10 per person per year, or 20 cents per person per week on each side of the transfer. So, the implied complaint of the average Northerner to the average African can be translated “I’ve been giving you 20 cents a week for years now, and you’re still poor – you must have squandered my generous help”.

This doesn’t answer the question of why Africa has remained poor while so many in Asia have grown rich, or at least much better off. But unless you impute truly magical rates of return to aid, the question “why hasn’t aid made Africans rich” can be answered very simply “there’s never been enough to make a difference”.

The “but we’ve given them billions” argument is a common one and it’s nice to see it neatly dismantled.

There is a counter-argument though; William Easterly Makes it in response to Steve Radelet here in a Council for Foreign Relations skirmish.

Radelet: …First, aid amounts have not been that big. Bill cites $2.3 trillion over fifty years, which sounds huge, but it translates to $46 billion a year, a modest amount for any global capital flow. Only about $26 billion per year went to low-income countries (as opposed to middle-income ones). This works out to be fourteen dollars per person per year in low-income countries—not exactly winning the lottery…

Easterly:…As for Big Pushes, the aid flows were certainly large to many of their recipients. The quarter of countries with the highest aid got on average 16 percent of their Gross National Income for the past forty-two years, year in and year out. Their average growth rate of income per head over that period: 0.4 percent per year…

16% of GNI per year! All of a sudden the numbers start sounding large again. At least for some countries.

Except that the countries that Easterly is talking about all have very low GNI per capita. (Heck that’s why we give them ODA right?).

So which figure is more salient: aid/GDP or aid/capita?

The answer to this, I think, hinges on how we actually might expect aid to help countries. If you believe in the old fashion big push theories, then aid/GDP will be most meaningful, as it will be relevant to savings gaps and the like. On the other hand, if you think aid will contribute to economic development through healthier, better- educated people then aid per capita is more relevant. Similarly, if you think aid shouldn’t be solely about economic development but that it should also simply thought of as a welfare enhancing transfer, aid/capita is the most meaningful figure.

Personally, I find classical big push type arguments very unconvincing (as does Bill Easterly, for what it’s worth), in which case it’s an open and shut argument: over the years, by the most meaningful measure, we really haven’t given much aid.



  1. It’d be helpful, I think, to define “aid.” What is it exactly? The presumption here is that “aid” is generally a transfer of wealth from rich to poor. I don’t know this to be true. I ask because “aid” in the aid portions of the US federal budget has equaled the transfer of weaponry or the subsidization of Americans working abroad or American agriculture. This counts as aid and it is likely, tallied in that “$500 billion in aid” to Africa figure. So before going into how we measure the impact of said “aid,” can we start be sorting out what “aid” is? And figuring out, really, is that aid?

    Comment by Carla Murphy — August 13, 2010 @ 11:08 pm

  2. I don’t know that aid-per-capita is even meaningful when thinking of “welfare-enhancing transfers.” If I give you a dollar, the welfare effect is contingent on whatever your current income is, right? We can’t compare aid-per-capita of one country to another and think about welfare without knowing the average income… and when we compare aid-per-capita with gdp-per-capita, all we’re really doing is looking at aid/gdp.

    Comment by Matt — August 13, 2010 @ 11:54 pm

  3. The level of the GNI (be it low as Liberia or high as Singapore) is irrelevant, I think. If either of them gets a consistent boost of %16 in foreign assistance, there should be noticeable rates of growth. In the African cases, there are not. I think I’d give this one to Easterly.

    Comment by Scott Gilmore — August 14, 2010 @ 1:11 am

  4. Thanks Carla,

    I agree: not all aid is equal. And a lot of what has been given over the years really doesn’t warrant the word aid. Which makes Radelet’s figure seem even smaller: $14/person/year and much of that was given poorly.

    One small caveat to the above though. The OECD DAC aid figures, which more of less everyone uses for overall aid totals do have quite stringent requirements about what qualifies as aid. And they exclude military aid (unless the soldiers in question are doing actual aid work). Nevertheless, a lot of other poorly given, ill-intended ‘assistance’ still qualifies.

    Comment by terence — August 14, 2010 @ 6:38 am

  5. Thanks Matt and Scott,

    Matt – I take your point: aid ought to have a bigger bang for its buck in poorer countries. So I concede the aid/gdp figure is meaningful in some sense. However, I still think that when weighing up aid vs the magnitude of the task we are asking of it, if you think (as I do) that aid’s most likely impact either on growth or on welfare is via education and health care, then aid per capita seems the most meaningful figure.

    Scott, more or less what I said to Matt: whether a 16% of GDP injection of aid should have a meaningful impact in a ldc or not, it seems to me, really depends on the mechanism you expect aid to do it’s work through: I think the most likely impact is via human capital hence my preference for per capita figures. That said, I’d still agree that aid of 16% of GDP ought to have some impact. But it will be small. Far too small for comments a long the line of “we gave all that aid and they’re still poor” to be valid arguments against aid.

    Thanks both for your comments.

    Comment by terence — August 14, 2010 @ 6:50 am

  6. To put it another way, Matt. According to Nation Master:

    PNG GDP in 2006 USD$5,653,885,000
    PNG Popn in 2008 5,931,769

    Fiji GDP in 2006 USD$2,821,606,000
    Fiji popn in 2008 931,741

    Do you really think say $20M dollars of aid is going to have the same or more impact on welfare in PNG than $10M will in Fiji?

    Comment by terence — August 14, 2010 @ 1:33 pm

  7. Terrence,

    My answer would be maybe, because you’re giving the same amounts relative to income. Knowing the income gives you vastly more information.

    Let me put it this way. Imagine two communities of equal population. Now imagine one is twice as rich as the other. Do you think X amount of aid is going to have the *same* welfare impact in both communities? Which is more informative, aid-per-person, or aid/gdp (which is the same as aid/income-per-person).

    Of course a million other factors matter (what the aid is, what and who it is spent on, etc). But if you’re really going to talk about welfare, which is an individual-level concept, you really need to be scaling it. A dollar given to me is not the equivalent to a dollar given to someone in Malawi, but by just looking at aid-per-person, that’s what you’re assuming.

    Comment by Matt — August 14, 2010 @ 9:33 pm

  8. Thanks Matt,

    That’s a thought-provoking response, and you’ve certainly made compelling arguments for why aid/GDP is relevant. Particularly, I think, why it ought to be relevant in guiding our decisions on where we give aid. Ceteris Paribas aid given to a poorer person or poorer country ought to have a higher welfare impact than aid given to a richer country (although, of course, as I think we’d both agree, it might be much harder to give aid well in a really poor country). However, when we are attempting sweeping evaluations of whether aid has worked or not – which is where this post started – I still think the fairest (albeit imperfect) metric for evaluating the impact of aid is vis a vis the number of people it’s impact has to be divided amongst. The reason I think this is because I think the primary conduit for either aid -> welfare or aid -> growth is through people, via health and education etc, rather than it’s direct impact on the economy.

    Of course, as you rightly point out, the welfare impact will vary depending on how bad aid is needed but when we’re looking at the total amount given to LDCs for example, which are all materially poor, I think this objection is negated somewhat by the partial welfare homogeneity across such countries.

    I definitely appreciate your comments though – thanks for helping me think more about this.


    Comment by terence — August 14, 2010 @ 9:46 pm

RSS feed for comments on this post.

Create a free website or blog at WordPress.com.

%d bloggers like this: