Aid, evidence and anecdotes

Tyler Cowen says that the debate about the effectiveness of foreign aid has improved in the last ten years. If so, then things must have been really bad a decade ago: it continues to astound me how many people are allowed to get away with peddling their prejudices without any meaningful evidence.

James Surowiecki has an interesting piece in the New Yorker which seems to support the case for aid:

Between 1946 and 1978, in fact, South Korea received nearly as much U.S. aid as the whole of Africa. Meanwhile, the billions that Taiwan got allowed it to fund a vast land-reform program and to eradicate malaria. And the U.S. gave the Asian Tigers more than money; it provided technical assistance and some military defense, and it offered preferential access to American markets. Coincidence? Perhaps. But the two Middle Eastern countries that have shown relatively steady and substantial economic growth—Israel and Turkey—have also received tens of billions of dollars in U.S. aid.

But this anecdotal analysis is no more valid than the opposite argument which was put by Bill Easterly in the New York Times on July 3rd:

From 1960 to 2003, we spent $568 billion (in today’s dollars) to end poverty in Africa. Yet these efforts still did not lift Africa from misery and stagnation.

Saying that we have given aid to Africa and yet Africans stayed poor is not an argument against aid; just as saying that we gave aid to Korea and they got rich is not an argument in favour of it.

As I argued here on June 30th, the question is whether aid makes a difference – and that requires some evidence about what would have happened in the absence of aid. You need to do a proper statistical analysis, controlling for other variables, to establish what difference, if any, aid makes to a country’s growth.

Plenty of studies have been done, and they nearly all find that aid is strongly, positively correlated with sustained economic growth in the medium term. My colleagues at the Center for Global Development did a study which looked at the relationship between aid and growth which finds:

higher-than-average short-impact aid to sub-Saharan Africa raised per capita growth rates there by about half a percentage point over the growth that would have been achieved by average aid flows. The results are highly statistically significant and stand up to a demanding array of tests …

And in a comprehensive survey of all the empirical research on thiark McGillivray at the OECD (pdf file here) finds that poverty would have been much higher in the absence of aid.


I’ve done a series of blog postings on aid effectiveness which set out the compelling micro and macro evidence for the effectiveness of aid. Jim at Our Word is Our Weapon is also a reliable source of evidence-based analysis.

7 thoughts on “Aid, evidence and anecdotes”

  1. I’ve just finished Easterly’s “quest for growth” and I didn’t get the impression he was anti-aid. I thought he was arguing that the wrong sort of aid has, in the past, been given in the wrong way, and that he supported that view with a lot of evidence to suggest that the affect of aid on growth has gnerally been extremely disappointing. Unless he was guilty of distortion in his use of statistics and data, it seems to me his case was compelling to say the least. Surely his position is just that we need to be smarter about aid?

    I don’t see why more people aren’t happier with a middle position – that sometimes aid does a great deal of good and that sometimes it doesn’t. I don’t understand why so many commentators seems to want to polarise their views towards either aid good or aid bad.

  2. Paddy

    It is true that some aid does not work, and some may do some harm. I don’t think even the most fervent proponent of aid would disagree with that. But aid that works well can be spectactularly effective.

    On average, aid has been shown time and time again to be strongly positively correlated with economic growth, which is a major driver of poverty reduction. In other words, the successful aid is far more important than the unsucessful.

    I am not sure what you mean by polarised views. Even accepting that some aid does not work, we have clear evidence that, on average, aid saves millions of lives a year. We could, easily and cheaply, save millions more by increasing aid spending and by improving the effectiveness with which it is spent. Our reluctance to do so is literally a matter of life and death. It is very frustrating that, in the face of such compelling evidence, some commentators continue to argue against an increase in aid and thereby contribute to avoidable deaths.

  3. I guess that by polarisation of views, I mean a tendency to needlessly overstate one’s own position, take the worst possible interpretation of the opposing side’s position, and selectively pick, evidence and arguments that support your own view adn ignore or diminish those that don’t.

    I get the feeling you’re not going to agree with me on this, but to my mind this tendency towards polarisation is visible all over the debate.

    Perhaps I ought to acknowledge my own lack of knowledge and shut the hell up, but this statement: “On average, aid has been shown time and time again to be strongly positively correlated with economic growth, which is a major driver of poverty reduction.” strikes me as an overstatement. Are there not also studies that show no correlation? I’m pretty sure I’ve read some. So, what do you do when you come across one? Immediately start looking for ways to dismiss it? If so, is that not a sign of polarisation? If (for the sake of illustrative argument) more than half of aid in the past has been given by World Bank bureaucrats looking to protect their budgets, prestige and cover their mistakes, to kleptocratic and incompetent states, then we would expect it not to produce growth and as a consequence for macro data to show no correlation with aid and growth. But why resist that argument? It’s perfectly consistent with the right sorts of aid saving lives, in a cost effective fashion, and is no argument “against aid”. Also, even if only 10% of aid does some good, and 90% does nothing – that would still mean “aid saving lives on average” – yet in such a world there’s clearly still plenty of valid criticisms to be made about aid, so why not embrace those criticisms rather than fight them?

    I guess, though, you might agree with much of the above – but that’s kind of my point – I get the impression that there’s not really as much disagreement as it appears between the likes of you and worstall, world-is-our-weapon and GDI, but just that the polarisation of views by both camps has the effect of aggravating matters.

    Anyway, I’ve tried to make this point before and it has fallen flat, so perhaps I’d better just leave it. And perhaps it’s just because I haven’t yet studied these questions enough myself that I haven’t hardened my own position, and figured out who’s wrong and who’s right.

  4. Paddy

    It is somewhat surprising to be accused of polarisation and partial representation of the arguments when my original post was specifically criticising the evidence base of an article (by James Surowiecki) even though he uses it to reach a conclusion that I agree with. That does not feel to me like the behaviour of someone who uncritically marshals every argument that agrees with my opinion, but rather someone who believes that arguments should be based on solid evidence.

    You asked “Are there not also studies that show no correlation?”. There are some, yes. But they are rare, they are almost always produced by researchers with a highly partisan agenda, and I cannot think of a single such study that is statistically robust.

    (Incidentally, I said that aid has been shown “time and time again” to be effective; this was intended to indicate that the vast preponderance of evidence points in this direction, but it is not intended to imply that there are no studies at all that disagree.)

    There are, as you say, plenty of criticisms to be made of aid, and I have made many of them here. I feel so strongly about this that most of my working days are devoted to trying to improve the way that donors give aid, so that it can be even more effective. But I base my analysis on evidence, not prejudices, and the evidence is very persuasive that countries that receive more aid on average grow faster than countries that do not. The argument for improving aid is not helped by claiming that it doesn’t work.

    There is a difference between being balanced and being objective. Balanced is giving equal weight to different views. Objective is basing conclusions on evidence. I plead guilty to not being balanced. I would not be balanced in taking account of the views of someone who believes in astrology, or creationism. They are entitled to their views; and I am entitled to take not no notice of them; I prefer to reach my views on empirical questions based on evidence. But I do aspire to be objective.

    To be honest, I do not know exactly what Tim Worstall believes about aid. But there quite widespread scepticism, especially among the commentarati, about whether aid works which I think should be confronted by those of us who have looked carefully at the evidence. If that feels partisan to you, then I’m sorry. But if we leave unchallenged the idea that the evidence on aid is ambiguous, the danger is that we will lose public support for increasing resources for aid, and that we will miss the chance to save millions of lives and improve the lives of millions of others. And I think that is important.

  5. Owen please do not take what I have written as personal criticism, even if I have expressed myself badly I have been trying to argue that there is a tendency towards needlessly drawing battle lines, misrepresenting the views of perceived opponents and taking more extreme positions than necessary etc. by commentators on all sides of the issue. It’s just the impression I’ve got by reading posts on various blogs, not from yours in particular. Perhaps that’s just what happens when people start arguing with each other. This is not quite the same thing as saying that you are wrong to take sides, or that the side you have taken is the wrong one. Your responses always make me feel like I’ve made a personal attack on you, which was never my intention (even if I did try to pick an example from you post to try and illustrate what I meant). Anyway, enough on this particular topic. I will continue to read what you’ve got to say with interest, but keep my observations to myself

  6. Paddy – please don’t keep your observations to yourself. I enjoy them: they are stimulating and they make me think. In turn, I am sorry if I sound defensive. Owen

Leave a Reply

Your email address will not be published. Required fields are marked *