Aid can promote growth, but don't rely on it to do so

This is the message I take from the most comprehensive analysis to date of the cross-national evidence on the effect of aid and growth. The paper in question is by Raghu Rajan and Arvind Subramanian, both formerly of the IMF. (Rajan is now back at the University of Chicago, and Subramanian is at the Peterson Institute of International Economics.) What sets the paper apart is its careful consideration of causality issues, and its broad scope and coverage, looking at the relationship under a number of different circumstances and settings.  Their bottom line:

we find little robust evidence of a positive (or negative) relationship between aid inflows into a country and its economic growth. We also find no evidence that aid works better in better policy or geographical environments, or that certain forms of aid work better than others. Our findings suggest that for aid to be effective in the future, the aid apparatus will have to be rethought.

Comments

[Per Kurowski]

Aid could work as those extra small wheels you put on the bike for your child to get a feel for it but, have them on too long, and he will start losing the necessary confidence to bike on his own, without wheels, and which of course is what development is all about

[robertdfeinman]

Sorry the quote reveals an internal contradiction. If there is no correlation between aid and development then why would restudying it change anything?

"we find little robust evidence...that certain forms of aid work better than others.
...
the aid apparatus will have to be rethought"

Or perhaps (given the author's affiliations) "rethought" is a euphemism for eliminated.

[Greg Sanders]

I'll have to read this in greater detail. Although even if there isn't a relationship to growth, there could be a rationale for aid as a means to alleviate human suffering. It might be worth applying the studies methodology to other statistics like lifespan, population in absolute poverty (although absent growth that would just mean income redistribution), or the like.

I'm also curious if they differentiated between disaster/reconstruction aid and other types of aid. On a brief skim they mentioned disasters as a possible cause of aid but don't seem to have otherwise addressed that issue.

[inthemachine]

"I'm also curious if they differentiated between disaster/reconstruction aid and other types of aid."

One of the trickiest elements to consider when looking at multi-generational aid data, is the effect of the cold war (and its attendant political incentives) on who did or did not receive aid. Considering the Cold war ended in the early 90's and the "war on terror" generated a new set of "priority countries" by 2002, we really only have 10 years worth of data that wasn't part of a larger, politically motivated global aid strategy.

[Roy Bland]

oh good gravy, I can't stand it.

I thought Roodman and/or Clemens et al had sorted this question out.

sigh.

better read it.

[Gillian]

Aid is increasingly tied to improved governance. For example Denmark supports the Tanzanian govt budget in three-year commitments with milestones tied to specific governance actions. Last year, Tanzania lost $2.5m of the Danish funding cos it was late in passing new anti-corruption legislation. Perhaps the 'rethinking of aid' will be in this area - where the aid is used to motivate better policy and institutions. The paper notes 'there are robust findings in the cross-country growth
literature for the importance of institutions and policies for growth'.

[dale]

I appreciate Greg's comments above. Economists may be interested in the relationship between aid and growth. But there are other reasons to provide aid.

[Ali Sohail (pakistan)]

well pointed out by Robert, but i feel that is a matter of getting lost within words!
Data, that i am exposed to tells me that certain forms do aid,enhance development and growth while others dont.
The period within which we measure effectiveness plays a pivotal role aswell.
Cheers

[Diego]

If aid within developing countries were directed to backward or peripheral regions/sectors, wouldn't one expect such aid to have a relatively low impact on aggregates? Still couldn't such aid have an important potential contribution within the limits of its scale?

[Justin Rietz]

Preface: Dani, I believe it would be interesting and useful to have a more general discussion about the validity of econometric models. Many of the papers to which you link are heavy on econometrics, so I it would be fair to explore theory vs. fact.

That being said....

I agree with the general conclusion of the paper. However, I think the methodology of the paper itself is flawed.

The use of econometric models to measure something as complex as the affects of foreign aid on economic growth is more or less futile. Even though the authors make a conscientious attempt to capture what they believe to be the most significant independent variables, it is impossible to include, with any amount of accuracy, factors that cannot be measured - factors that tend to be the most influential.

Can an economist accurately model the actions of millions of people across entirely different cultures with different histories, different religions, different natural resources and different cultural norms – not forgetting that the model must take into account the interaction of these variables? While proxy measures may be used, it is best to avoid doing so when the model is (or should be) highly dependent on the actual variable. It isn't surprising that every economist who takes on similar research gets dissimilar results. Given the complexity of the issue, I think this is the only thing we could accurately determine statistically!

The authors try to take into account each country's political and social backdrop, but the data is questionable. The inclusion of variables for whether or not a nation has been a colony and something called "Institutional Quality" (which should have been explored more in the paper) may in fact _weaken_ the model as there is no mathematical way to know if these statistics accurately represent what they claim.

Ultimately, modeling of such topics detracts from more useful endeavors. The researchers' time would have been better spent studying the politics and histories of each country, and interviewing people at ground zero. They may then have been able to come up with useful, anecdotal reasons for whether or not foreign aid works.

[scott bayley]

I agree with Dani Rodrik's view that the paper looked at things from a number of different circumstances/settings but disagree that the authors undertook took a careful consideration of causality issues. I think they encountered many of the typical problems that have been identified in the literature. For example causal analysis using statistical models is very problematic, see:

Glazerman et al, 2002, 'Nonexperimental replications of social experiments: A systematic review'. Mathematica Policy Research Inc.

http://www.mathematica-mpr.com/publications/PDFs/nonexperimentalreps.pdf

The authors also violated a number of the statistical assumptions underlying the technique of multiple regression. MR results are very sensitive to violations of assumptions (e.g. ommitted variables, irrelevant variables, non random sampling, non linear relationships, interval level measurement, homoscedasticity, etc). These assumptions are discussed in the following website article:

http://www2.chass.ncsu.edu/garson/PA765/regress.htm

Finally, to test for causality 3 types of evidence is required:

1) Covariation. The nature of the relationships that we analyze after our observations are made. We look to see if the dependent and independent variables are related as expected.

2) Temporal order. By employing different kinds of research designs, evaluators can exert influence over temporal order. (i.e. cause occurs in time before the effect)

3) Elimination of rival hypotheses. This can also be treated in the design of research projects, if the competing explanations are incorporated as variables in the design.

In the article in question the authors focussed on type 1) evidence (although their sample sizes were much too small); partially dealt with 2); and failed to come to grips with 3). Hence their conclusions about the lack of effectiveness of aid are not justified. The application of these 3 criteria for causal analysis is discussed at length in Cook and Campbell, 1979, 'Quasi-experimentation: Design and analysis issues for field settings'; and in Shaddish, Cook and Campbell, 2002, 'Experimental and quasi-experimental designs for generalized causal inference'.

[rolex watches]

People usually say :"Seeing is believing." http://www.tt88times.com
Each attempt has a corresponding gain, in part or obvious, or vague. At least we have the kind of satisfaction After I bought this watch ,in a sense,it means a great deal to me. http://www.fashionhairfu.com

[ugg classic knit]

People usually say :"Seeing is believing." http://jordanmass.com Each attempt has a orresponding gain, in part or obvious, or vague. At least we have the kind of satisfaction After I bought nfl jersey ,in a sense,it means a great deal to me. http://www.fansshirt.com

[ugg classic knit]

People usually say :"Seeing is believing." http://jordanmass.com Each attempt has a orresponding gain, in part or obvious, or vague. At least we have the kind of satisfaction After I bought nfl jersey ,in a sense,it means a great deal to me. http://www.fansshirt.com

[nfl jersey]

People usually say :"Seeing is believing." http://jordanmass.com Each attempt has a orresponding gain, in part or obvious, or vague. At least we have the kind of satisfaction After I bought this watch ,in a sense,it means a great deal to me. http://boon-shoes.com

[nfl jersey]

People usually say :"Seeing is believing." http://jordanmass.com Each attempt has a orresponding gain, in part or obvious, or vague. At least we have the kind of satisfaction After I bought this watch ,in a sense,it means a great deal to me. http://boon-shoes.com

[jordan shoes]

People usually say :"Seeing is believing." http://jordanmass.com Each attempt has a orresponding gain, in part or obvious, or vague. At least we have the kind of satisfaction After I bought this watch ,in a sense,it means a great deal to me. http://boon-shoes.com

[jordan shoes]

People usually say :"Seeing is believing." http://jordanmass.com Each attempt has a orresponding gain, in part or obvious, or vague. At least we have the kind of satisfaction After I bought this watch ,in a sense,it means a great deal to me. http://boon-shoes.com

[Nike Air Jordan]

The good news, thank you!

[lachou]

I like these articles , democratic countries formulate policies are both team and the interests of the people of power, but the undemocratic countries rarely for the interests of the people, http://www.jerseysky.com it is the sorrow of many democracy, I like the western and northern state policy.
http://www.cheapsaleing.com
http://www.nikeairmaxshoe.ocm

[cheap jordans]

I think this is a great post. One thing that I find the most helpful is number five. Sometimes when I write, I just let the flow of the words and information come out so much that I loose the purpose. It’s only after editing when I realize what I’ve done. There’s defiantly a lot of great tips here I’m going to try to be more aware of.

[jordan shoes]

It's good to see this information in your post, i was looking the same but there was not any proper resource, than now i have the link which i was looking for my research.

[Retro Jordans]

Thank you for that great tutorial. I think it is the only one on the Web. The three parts with the picture are in my files, now.
You make me happy :-)

[أيفون]

this is a very good analyzing i also agree with this

[Account Deleted]

raf

Thanks so much.

[Account Deleted]

Wholesale Beads we find little robust evidence of a positive (or negative) relationship between aid inflows into a country and its economic growth. We also find no evidence that aid works better in better policy or geographical environments

[Account Deleted]

Raf
raf

Thanks for this ...

[Account Deleted]

Many places and centers offer business and trade promotions to both buyers and supplier.What about the differences in skill intensities across industries? The job losses in the relatively unskilled-labor intensive battery industry should have little effect on the relatively skilled-labor intensive machinery
sexshop
sexyshop
sexshop online

[Account Deleted]

This site is excellent and so is how the subject matter was explained. I also like some of the comments too.Waiting for next post.

organic seo service

[Account Deleted]

recovery powerpoint file

Thanks for the interesting article!