not knowing whether aid works vs knowing it doesn’t work

The research exploring the relationship between foreign aid and economic growth has shown a fragile relationship (when it has shown a relationship at all!).  I just enjoyed Aid Effectiveness – Opening the Black Box, by Bourguignon and Sundberg (2007), which gives a nice summary of why we shouldn’t assume that means that aid doesn’t work.  (Rather it means that either aid doesn’t work or we’re measuring aid wrong or we can’t capture how long aid should take right or …)

It is no surprise that reduced form analysis shows tenuous links between aid and development outcomes. Aid has often been for non-developmental objectives, such as disaster relief or for military and political ends. Much aid is lost due to instability and conflict: roughly half of aid to Sub-Saharan Africa has gone to countries facing civil war and/or frequent military coups (Fitzpatrick et al, 2007). Much (though not all) aid has also been wasted on poorly conceived and executed projects and programs, often fettered by debatable conditionality. And from a statistical point of view many technical problems arise: distinguishing short vs. long term impact, problems with endogeneity of the aid-growth relationship, difficulty determining the direction of causality or controlling for country-specific characteristics, etc, (Bourguignon and Leipziger, 2006).

Case studies do not avoid this due to the difficulty of establishing the counterfactual: Easterly (2006) argues that aid is not associated with growth in Africa, whereas Collier (2006) argues that in the absence of aid growth would have been far worse. Moreover, the multi-dimensionality of development objectives-mean income, poverty, literacy, access to sanitation, inoculations-further complicates empirical analysis.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s