Aid Thoughts

Yemisrach Kifle at Change.org's Global Health blog argues that the debate over whether aid works does little to distinguish different types of aid, some of which she claims are effective:

...to simply declare aid as a whole an ineffective poverty reduction tool runs counter to the facts on the ground. It also ignores the people who have benefited from the hand up international aid has provided. Ask the residents of the small Kenyan village of Sauri, one that's part of Columbia University professor Jeffery Sachs' Millennium Villages project. A recent report by the New York Times recounts the visible difference international aid has made in the lives of the people from this small village.

I think there are two broad, reasonable metrics for judging whether or not a given subset of aid "works:"

1. Does the expected future level of the recipient's welfare increase, conditional on there being no more aid in the future? (i.e. does the aid help create or sustain a growth path in welfare, health, income, etc). Another way to think of these is: if I give someone aid in year T and only in year T, will their welfare in year T+X be be higher than it would have been without the aid? The larger the value of X that this is true for the more effective I'd consider the aid. Think of scenarios where this number is negative.
2. Given metric #1, will the next dollar spent on aid in a given context be more effective than a dollar spent in another context? (What is the opportunity cost of this aid?).

This is actually a broader definition than it seems.  Depending on the context, non-governmental aid which directly provides education and health services "works", although likely only for a limited X. Investments in education and health systems, however, are going to be more effective by this definition, even if they mean sacrificing more immediate welfare gains.

Does the Millennium Village model "work" by this metric? Do we have enough evidence yet to even answer this question?

Some argue that aid is not about sustainability at all - that we're just reducing suffering while we wait for developing countries to pick themselves up. For metric 1, this is still acceptable as long as the expected future welfare gain is positive, but what if it isn't?

A few nights ago Jon Stewart, the host of the satirical news programme the Daily Show, interviewed Marc Thiessen, a proponent of waterboarding as a legal method of interrogation. Thiessen offered up a number of examples where terror suspects had been waterboarded and had produced intelligence that led to the disruption of terrorist plots.

Stewart tried to explain to Thiessen that positive examples weren't enough to establish certainty in the causal relationship: that just because waterboarding was used to discover this intelligence, it didn't follow that waterboarding was the only way the authorities might have discovered that intelligence.

Aid advocates often make a similar fallacy in analysing the historical impact of aid (and no, I'm not trying to make a connection between aid and waterboarding). They point to some obvious wins - usually education, health and agriculture. Kifle makes the argument:

Has [aid] failed when it enables that child to attend a school that has just been built by an international fund?

These sort of arguments ignore the more crucial questions: were there any other paths we might have taken to arrive at the same outcome (educated children?) To me, arguing that aid accomplished that goal isn't enough - advocates need to make a compelling case that aid was the best way to meet that goal.

This brings us back to metrics 1 and 2.