Polman, Collier, and the Rational Actor Model
Aidwatch has another interesting post up, featuring David Zetland and his review of the review of Linda Polman’s new book “Crisis Caravan“. I made a snide remark about it on twitter, and Bill Easterly rightly suggested I take it up with the author not Aidwatch. So…..
Polman is a Dutch journalist who is writing about the aid and humanitarian industry. As Zetland explains:
The central thesis of this book (as presented in the review) is that the people who deliver aid are addicted to horror stories and starving kids, and this addiction is fed by those who benefit from aid, whether they be local leaders, militias committing atrocities or even victims who don’t wear their prosthetic legs because they can get more attention with their stumps.
I watched Polman tout the book on the Daily Show, and went to her launch at the Half King in NYC a couple weeks back. There are three main problems with it:
- Despite what Zetland says, anecdotes are not data. The book reminds me of “Lords of Poverty“, which was shooting in the right direction, but relied almost solely on entertaining anecdotes. The aid industry is a total mess largely because huge decisions are made on what feels right, truthiness, if you will. In this sense, Polman is right in general, the aid industry is a mess, but her anecdotes focus on coordination and motivation issues, not on aid effectiveness. Polman’s got lots of “truthiness”, but little else.
- One of her prescriptions is some massive European-style coordinating and governance body which would oversee all NGOs around the world and ensure that they have ethical standards. This would only make matters worse, not better. If you ever sat through an interminable UN Country Team coordination meeting in a place like Afghanistan or Haiti, you know of what I speak.
- Her argument on the motivation of warlords reminds me of a ridiculous formula Paul Collier once used to explain the decision making process of rebels (see below). It combined the likelihood of victory, the taxable revenue at stake, the size of the population, and five other factors to arrive at the “Rebel Utility Function”. Very impressive, but the only problem is that I’ve yet to meet a rebel leader who subscribes to the “rational actor model”. Most of them are more than a little nuts, and even the sane ones take decisions based on a much more subjective basis such as how angry they are, and how many guns they have access to.
To sum up, bad book, bad ideas, and I disagree with David that anecdotes equal data.