Aid, Dis-aggregation, and Infant Mortality

Easterly primarily argues that aid has failed to promote growth. His evidence seems relatively convincing. However, GDP growth is not the primary goal of most aid projects. Aid is typically designed to improve education, health, infrastructure, or improve governance, all with the goal of alleviating poverty. While poverty is very clearly linked to GDP and GDP growth, this may be through the long term. Furthermore, GDP could grow in some parts of the country, leaving others impoverished. Easterly is jumping to the end goal but doesn't look methodologically at where in the chain the failure is occurring. He suggests doing this through more micro-level assessments of what works. However, I think that more could be done at the macro-level to see whether aid is failing and why as well.  In order to assess why aid fails, I think Easterly could have constructed a model linking aid to GDP growth and test each link in this chain.  The model should be (un-academically and non-mathematically expressed because it would be a series of equations with other factors included) something like this:

(created by me, ideas partially taken from Fidley and Hawkins (2009))

It would be useful to look at which of these types of aid is failing to create outputs in the "quality of life" section and then which of these outputs is not strongly linked to poverty reduction and growth outcomes, thereby accounting for the lack of evidence that aid works. If the research showed that aid is a significant contributor to improved health, reduced conflict, improved education, improved investment, and democracy, all measured separately, then aid could be taken as a useful tool to alleviating poverty. The question would then be, why are improved health, education, governance, etc. unsuccessful in improving growth rates or what is wrong with the literature.

Since this would be a massive undertaking, I decided to look at a portion of this: whether health aid reduces infant mortality. In contrast to the effects of general aid on growth, health aid seems to also have a controversial  impact on reducing infant mortality. Mishra and Newhouse (2009) uses data from 118 countries between 1973 and 2004 in a dynamic panel with fixed effects to show that doubling health aid is associated with a two percent drop in the infant mortality rate, implying that for the average country, an increase in per capita health aid of $1.60 per year is associated with 1.5 less deaths per thousand births. In contrast, Boone (1995) finds no impact. Thus, from a cursory look at the research, it seems that health aid is a part of the problem. If health aid is not consistently found to improve health outputs, it could not reduce poverty or increase growth outcomes.

Regardless, before believing Easterly's conclusion that aid is generally ineffective and aid agencies must be held accountable and more careful measurement of each intervention is needed, I would want to make sure that the careful measurements are being done efficiently. More accountability to aid agencies means more accountability for those who they fund. This would direct the accountability upwards to the donors as opposed to the poor people in the local communities they serve. Thus, more accountability for aid agencies could mean less accountability to local people. Overly extensive date reporting standards also result in inefficient aid - for example, the requirements for reporting for EU aid take about 50 percent of the total time and thereby expenses of the project. Furthermore, requirements for types of projects that are proven to be effective could reduce innovation at the local NGO level, who are more likely to be seeking solutions to local problems than planning grand schemes from above.