Making Aid Work

My op-ed last Tuesday was on the effectiveness of foreign aid, looking especially at AusAID’s efforts (via its Office of Development Effectiveness) to ensure that the multi-billion dollar increase in foreign aid over the coming years isn’t wasted. Full text below. As seasoned readers will guess, I was itching to use the phrase “they should do more randomised trials”, but managed to hold myself back.

Sorting Good Aid From Bad, Australian Financial Review, 20 May 2008

In his 2005 book The Undercover Economist, Tim Harford revealed that fair trade coffee in some London stores cost an extra 25 pence per cup, but that only about a tenth of the markup actually reached the coffee farmer. A fair trade cappuccino might have given the drinker a warm inner glow, but it didn’t do much to reduce world poverty.

Its harshest critics sometimes say the same about overseas aid. They point out that unlike the voters who judge domestic programs, the recipients of foreign aid cannot punish bad policy at the ballot box. Aid’s critics say that with a few exceptions – like disaster assistance – the world would be better off with less aid, not more.

With this year’s budget promising a major increase, Australia’s overseas aid is set to rise from 0.3% to 0.5% of national income by 2015. So the billion-dollar question is: does aid work?

At an aggregate level, it turns out to be depressingly difficult to identify an impact of foreign aid on a country’s level of development.  In the 1990s, a series of studies reached ambiguous conclusions. In a famous paper in 2000, Craig Burnside and David Dollar found that that aid did raise growth – so long as the recipient country had good fiscal, monetary and trade policies. Three years later, William Easterly, Ross Levine, and David Roodman found that these results were fragile, and did not hold up when more years of data were added to the analysis. Subsequent research by Paul Collier contends that even in badly-governed Africa, poverty rates would today be much higher had the continent received no aid.

Most recently, Jeffrey Sachs has argued that much foreign aid in the post-war era was directed towards winning the Cold War rather than helping the poor. According to Sachs, we should not discriminate against countries with bad institutions; instead we should be ruthless about not giving the wrong kind of aid. Good aid, argues Sachs, can make a difference.

This research offers two challenges to Australia. If aid only works when policies and institutions are strong, should we keep giving around one-third of our aid to Papua New Guinea and the Pacific? Ethically, it would be tough to justify withholding charity from some of the world’s neediest people. And given that they are on our doorstep, it is difficult to imagine excluding ‘fragile states’ from the Australian aid program. Yet the research suggests that despite all the hard work that is being done by aid workers in these countries, there is little chance living standards will take off any time soon.

The other challenge for our aid program is to distinguish between good and bad aid. This is a particularly thorny question at a time when the program is being expanded. This year, AusAID’s budget will rise from $3.2 to $3.7 billion. The last thing we want to happen is for the agency to lower its standards in order to get more cash out the door.

Here, the brightest hope for the aid program is that AusAID is taking seriously the question of aid effectiveness. On the last parliamentary sitting day before Easter, Parliamentary Secretary Bob McMullan quietly tabled the first annual report of the Office of Development Effectiveness – an agency established to monitor the overseas aid program.

Carrying out what it called a “health check of the Australian aid program”, the report critiqued aspects of the aid program in surprisingly robust terms. Discussing technical assistance (providing advice, or sending an Australian public servant to fill a role in a developing country bureaucracy), it pointed out that we spend twice as much as other rich donors, and warned that some of our technical assistance programs may undermine the capacity of poor countries to govern themselves. Discussing evaluation, the report emphasised the importance of carrying out more impact evaluations – asking not only whether the program was properly administered, but also whether it helped improve the lives of the poor.

The frankness of AusAID’s Annual Review of Development Effectiveness stands in stark contrast to the self-congratulatory mush of most government reports. Indeed, it is impossible to imagine the health, education, or defence departments putting out such a self-critical report. This kind of hard-headed approach is particularly valuable given that overseas aid is an area where too many advocates have focused on inputs (aid as a percentage of national income) rather than outputs (poverty reduction). Alongside a multi-billion dollar increase in foreign aid must come more scrutiny of results. If we care more about making a difference than getting a warm inner glow, distinguishing good aid from bad aid is the only way to go.

Andrew Leigh is an economist in the Research School of Social Sciences at the Australian National University.

Update: RMIT’s Simon Feeny directs me to his terrific survey paper on aid effectiveness, available in working paper form here.

This entry was posted in Trade & Development. Bookmark the permalink.

11 Responses to Making Aid Work

  1. Kevin Cox says:

    Andrew might I suggest that rather than beat your head against a brick wall with your calls for randomised trials you approach the problem by thinking of better ways to

    1. work out ways to better measure and specify outcomes of any program.
    2. collect more and better information from any program that is put in place.

    If this is done then you will have information that may preclude the need to set up formalised randomised trials and better still you will have systems in place that will see what happens over time and see the results of changes made as a result of analysis.

    That is think of ways of turning programs into a continuous “randomised” trial.

  2. Brendan says:

    Wouldn’t free trade be so much better for the world’s poor?

  3. Matthew says:

    Brendan, I don’t think that it’s an either/ or situation. We can reduce tariffs and give aid (the good kind). Both will help.

  4. conrad says:

    Given the small sample sizes you are likely to have and the huge variety in the way aid is given (and huge variance amongst recipients for that matter), I find it hard to imagine how you would ever run a randomized trial for many areas of foreign aid.

  5. Mike Pepperday says:

    I think institutions are important.

    Aid has its hand-out and its hand-up aspects. Sending Australian administrative officials might undermine development of domestic expertise though it would presumably be intended to train domestic expertise.

    A former senior administrator of such training programs in the Solomons gave a most interesting university seminar about a year ago. Australians were in there training the locals in how to run a modern country. She mentioned in passing something about the electoral system and I asked her later about it. She said they were leaving it alone as the effects of change are too unpredictable.

    I asked then if she was aware that if the the Australian government effort succeeds, the Solomons will become the world’s first successful country that has a single chamber of parliament made of single-member electorates. She was not aware and seemed a bit shocked.

    If I had wanted to rub it in I could have asked whether our model in the Solomons is Mauritius. Or was it pre-1996 New Zealand? Surely it would not be 20C Northern Ireland!

    Institutions are important. The Solomons, PNG, Vanuatu and Tonga have no chance.

  6. christine says:

    Mike P: Surely the model is Queensland? Admittedly, not a completely independent country, but what better model could there be for a political system? What, you think abolishing the upper house did anything but allow the government to run the economy properly for once? (Note: tongue in cheek.)

    Ooh also, Canada doesn’t have an elected upper house, and the Senate is pretty much completely useless. It would, I think, count as a successful country with a single (elected) chamber, all of whom come from single-member electorates. But this is not my area of competence, and I’m not raising it to dispute your point.

    More importantly, perhaps, having advisors from Australia try to even suggest such serious political reform might be a bit on the nose, mightn’t it?

  7. Mike Pepperday says:

    I have long thought that the reason Queensland doesn’t go completely bananas is that it is a sub-polity and, when necessary, Sydney can send Four Corners up there to sort them out. BTW it’s a great story, Ted Theodore’s “suicide squad” abolition of the Qld upper house contrary to the outcome of a referendum he held. (And another saga how in NSW the squad reneged, ca 1930.)

    Canada’s upper house is a disgrace. The Lords is not much better. But if you have single member lower house then anything – anything- is better than none. NZ’s was as big an abomination as Canada’s (appointed sinecures) and they abolished it by suicide squad about 1950. At around the same time Denmark and Sweden did away with their upper houses but whereas those two countries forgot about them in a week (they are PR polities), in NZ they couldn’t do without it. Eventually (1996), and partly by good luck, NZ took up PR. NZ is set for the future.

    In Northern Ireland they tried to introduce PR in 1982 but after 60 years of majoritarian polarisation it wouldn’t stick. They tried and failed again in 1998. Third time lucky it seems. Note that they knew, at least since 1982, what they had to do. Continental Europe knew in the 19C.

    Yes, Australian advisors telling them what to do would be on the nose. It is not going to happen. I have been watching the Pacific watchers and I cannot see that they have any interest in seeing the problem solved. As for Ausaid, my impression is of a bunch of touchy-feelies who would be far too sensitive of those countries’ feelings to be so crass.

    Those countries are on course for civil war. The little ones are of no consequence – only a few hundred thousand people and we can wring our hands and make the right noises – but PNG is six millions and very rich in resources. When the coup occurs, people there will be immensely relieved at finally getting some order. Bougainville will be ethnically cleansed, land tenure will be modernised, and Queensland will have half a million refugees.

    But at least we won’t be on the nose.

  8. Andrew Leigh says:

    Conrad, the proposal for randomised trials isn’t about randomising whole programs, it’s about evaluating individual programs by randomising across individuals (analogy: we don’t randomise drugs on the PBS, but we do randomised trials to decide which drugs get approval). The MIT Poverty Action Lab has some good examples.

  9. Pingback: AusAID Technical Assistance: Not Sexy but Still Attractive « East Asia Forum

  10. christine says:

    Mike: really interesting comments, thanks. And I see your point that actually trying and failing and being seen as paternalistic bastards could be better than the alternative.

  11. Angus says:


    Spot on with your Pacific Islands comments, both about the coming civil wars/coups and the touchy-feely uselessness of AusAid. And a very, very good point about Queensland having half a million new residents when the PNG coup does come – if many thousand boat people could make it here from Vietnam a quick hop from PNG is going to be no problem at all. They could float over on some of the logs left lying around from the total destruction of the rainforests that would follow about 10 seconds after the coup.

Comments are closed.