Brad DeLong excerpts a white paper on financial sustainability that immediately made me think of the failed policies of NGOs in Haiti.
bq. While sustainability is certainly a desirable goal, it may be difficult to achieve. Teaching people to fish rather than providing fish is great if it works, but this method works only if the donor knows more about fishing in the local area than the people who live there, and only if the donor can transfer this knowledge. Yet it is difficult for outsiders to understand how institutions, politics and societies function, let alone how to influence them in a way that does not create unforeseen consequences. Even if a hypothetical planner could target foreign assistance so as to change communities and institutions for the better, the principal-agent problems involved in foreign assistance make it hard to do this in practice. It is difficult enough to monitor aid workers handing out fish, since they are not subject to market pressures, nor held democratically accountable to the people who they are charged with serving. However, at least one can determine whether fish have reached the intended recipients, and presume that if so, the recipients are better off. In contrast, it is much more difficult to determine whether training sessions for leaders of non-governmental organizations (NGOs) working with the local fishermen have in fact made anyone better off. Foreign aid workers may provide encouraging anecdotes, but given their incentive to select among anecdotes, it is difficult to know whether donors would have been better off simply handing out fish.