This post is a little different, as it is my exchange with David Roodman over a paper I co-wrote, and the selectivity part of the Commitment to Development Index. He has written some of my favourite articles in development economics, and it felt odd to disagree with him. But disagree I did. David kindly agreed for me to reproduce our conversation here: it hasn’t been edited apart from one minor addition by David (for clarification). I’ve published it because I think the discussion might be interesting for others that are thinking about the whole host of indices that have been published in recent weeks.
Some context: Measure for measures is paper I co-authored with Edward Anderson that axiomatically works its way through various measures of donor allocative performance: how a donor divides its aid between different recipients. The core finding is that no measure currently on the market is always sensitive to changes in a country’s policy, poverty and population. David is behind the Commitment to Development Index, of which we discuss the selectivity part. It is completely insensitive to population: thus it values aid given to a large and small recipient equally. We felt this was a major flaw in a selectivity weight, he doesn’t.
A disclaimer: we conducted our exchange on google+ which is not exactly an ideal place for detailed academic debate and neither of us had the debate expecting it to be made public. So these comments may not perfectly represent our respective views. It is long, so its below the fold…
