Andy Sumner and I are working with some excellent research assistants at IDS to complete a number crunch on the bilateral spend in sub-Saharan Africa and Asia (we thought we could do it quickly, but it is actually quite difficult to ensure that pre and post BAR numbers are comparable! DFID transparency gurus, please note).
What the numbers are showing so far is: (a) the spend is about as pro-poor as before, which is very good news, but (b) the spend is higher in countries where freedom scores are lower (Freedom House index) and where corruption scores are higher (Transparency International index).
The freedom and corruption outcomes are no big surprise since definitions of fragility (see Stewart and Brown) relate to state authority failures (the state cannot protect its citizens against violence and criminality), service delivery failures and legitimacy failures (the state has very little support from its citizens).
If we believe these indices, this means that aid does not only have a double responsibility, as Secretary of State Andrew Mitchell puts it, but actually a triple responsibility: make sure aid: (a) has a transformational impact on people's lives, (b) does this with an emphasis on value for money, and, vitally, (c) does not incentivise further closing down and corruption in recipient countries.
How to use aid for development in less free and more corrupt societies? An open question, and one that all of us should be trying help answer. I don't think it is about new forms of conditionality, but rather it's about the way that aid is used.
It seems to me that, whenever possible, aid should be directed to giving citizens "voice and choice". One obvious examples is cash transfers direct to households. These are set to expand within DFID and other aid donors. Another, less obvious but potentially transformative option is "people powered evaluation" -- citizens living in recipient countries delivering their verdict on the difference UKAid makes on the ground.
I do not have definitive evidence on whether it is practical or will make a difference (although IDS is currently involved in a randomised controlled trial of such an M&E system (versus a conventional M&E system) in the Philippines in the context of cocoa, coconut and rice production).
But if successful it would (a) give a more "grounded and rounded" view of the difference aid makes, (b) make stories of impact more convincing to UK citizens--being a counterpoint to those who argue that people in the UK defend aid more vigorously than people in recipient countries do and, crucially, would (c) promote openness, transparency and accountability in a way that puts pragmatism ahead of rhetoric.
If it worked, it could transform the aid debate by rounding out the transparency, accountability and results agendas and would build huge trust with the UK public.
In the context of cash on delivery it would be like giving sign-off authority on the delivery docket to citizens in, say, Tanzania, Ethiopia and Pakistan.