I think it could be the basis of a new consensus – linking a (broadly European) agenda of giving governments in developing countries more freedom to choose their own approaches to development with a (broadly North American) agenda of ensuring that aid is more firmly linked to results.
There are, as JMS rightly says, some issues that need to be resolved: how to define the outputs exactly, and what to do about countries that are unable to make progress towards results. As he says, the first can be solved. The second is of course a problem common with any system of payment by results: it defeats the point of results-based-aid if you get the aid irrespective of whether you deliver the results. My answer is that if (and when) we want to give aid to countries that are not able to produce results (and there will be occasions when we do) then we should devise a separate mechanism for doing so, with its own criteria. That is not a critique of Cash on Delivery Aid, but a case for ensuring that it is not the only system for providing aid.