My view is that we need a simple mechanism (like Cash on Delivery, though it needs to be tested) that funnels aid to governments that can use it well, with the minimum of hassle and interference from outside; and a different funding mechanism for governments that cannot, or will not, provide services to the poor. In the latter case, it is unlikely that providing them with a lot of money will do the trick – that’s rather like pushing on a piece of string. For them, the interventions might not involve a lot of money, but some combination of support for civil society and capacity building (though done well, unlike most capacity building in development). Giving a lot of money to governments that don’t deliver deprives the citizens of countries that could use the money of services, while doing little if anything for the citizens of the country that can’t or won’t deliver.

But let’s be clear. This is still mostly an untested concept. There are lots of devils in lots of details yet to be discovered. The key area is how you define and measure results. If not defined properly you still end up paying for processes (redefined as results) and not getting outcomes that matter. Reread your own blog on Targets, Owen, and all of those issues you rightly raise describe the risks involved in pay for performance.

There are other issues of attribution. Will donors feel good about paying for results when there is evidence that it had nothing to do with what the government or implementing agency did, but was just good luck? And the opposite case– what if the implementing agency made a serious, best practice effort but didn’t get the desired results due to factors beyond its control– war, environmental disaster, etc.

Jeff – I agree with this – especially that it is untested and that the devil is in the detail. I think we should test it. (Some of the points you raise are partly addressed in my response to the CAFOD brief.

I’m not sure this is terribly insightful, it’s just another way of describing/explaining the resistance to cash on delivery aid you report. I’m not sure whether one would say this tendency is so strong that you ought to give up on cash on delivery aid and go home, because the donors are never going to wear it, for this reason, or whether it simply explains how you have some barriers to over come, which can nonetheless be overcome.

There now follows a cheap point – what would your CoD education donors do if it turned out the recipients were effectively educating their young citizens in creationism, and that homosexuals ought to be executed? It’s not so easy to focus on effectiveness and leave policy to recipients, when policy is what you are being effective at.

This reminds me of the debate in the UK voluntary sector when NCVO (National Council for Voluntary Organisations) published a book on “Outcome Funding” from the US-based Rensselaerville Institute and invited the authors over to the UK.

Not sure what came of it all, but it certainly provoked strong views on both sides of the debate.

Looks from their website like it is still available:

http://www.rinstitute.org/publication_detail.php?prodid=2

James

I have seen issues with the measurement, as mentioned in the first comment. For example, GAVI takes a performance-based measurement approach and I’ve seen countries falsely inflate their numbers so as to secure more GAVI support. My personal observations are backed up by a Lancet study (http://www.aegis.com/news/ap/2008/AP081240.html). I’ve also seen countries falsely deflate their numbers to maintain eligibility for GAVI funds. I’m not blaming GAVI for this; they are well aware of the limitations of the funding approach.

I think April also makes a good point. If we look at education reform in the US, we see a good example of what she’s talking about. The Bush administration instituted a pay-for-performance approach by funding public schools based on standardized test performance. The result is that teachers teach to the test rather than teaching students to be critical thinkers. Another result is that courses that don’t directly affect test scores (e.g. music, health) are the first courses to be cut, despite their importance to developing well-rounded, productive members of society.

I’m not saying that performance-based aid is a bad idea. I am saying that we should think about how the mechanisms provide incentives and what they provide incentive for.

Your post raises some interesting points about why COD may be seen as a threat to a few people who work in government donor agencies and raises important points about incentives for such staff. However, to suggest that individuals within CAFOD are feeling these threats is to misunderstand the work of CAFOD.

We are a partnership agency. We do not implement programmes directly but work with local civil society, particularly agencies of the Catholic Church. A large part of the work we do is to help these groups hold their governments to account so a move to COD aid, if it works as you say, could help support this work. I don’t think that roles within CAFOD would change significantly if this system was adopted.

A discussion on incentives is a valid dimension but CAFOD’s arguments should be taken as a contribution to the debate rather than seen as motivated by this.

Thanks Amy. I did not mean to imply that you personally, or anyone at CAFOD, felt threatened by results-based aid. I have never discussed results-based aid with you or anyone else at CAFOD so I would have no basis for thinking that. But the issues raised in your paper – which are consistent with some of the concerns expressed by other people in other forums – mainly boil down to the question of whether developing countries should be able to decide for themselves how to deliver services and to whom, and be accountable for that; or whether donors should continue to try to intervene in that process. Those are concerns that I hear raised by development professionals whose work might be affected by moving to results based aid in the way I describe. (They are people in NGOs as well as government agencies. I have heard several people in NGOs oppose results-based aid because the government, given a choice, might not choose to deliver services through them.) It is the motives and interests of these people that I am talking about, not yours. And, as I hope I made clear in the post, I think very highly of people who work in development, and there are many people who would take a very different view from the generalisation that I have presented here.

Who can oppose rewarding outcomes? BUT measurement is tricky/dodgy even in the UK & reliable measurement requires democratic accountability and feedback and a lot of transparency. Not always present in developing countries. This is not mentioned in your blog, neither by CAFOD -I think.

I feel like a reactionary and it does not feel good. However, my long and tedious work on performance management (including UK govt. targets) and on data (as a little boy in the Central Bank of Peru) leave me no option but to be a bit skeptic of cash on aid delivery without democratic accountability.