Does anyone really think that giving aid to Robert Mugabe would do much to increase the long-term growth prospects of Zimbabwe? Conversely, does anyone really think that money well invested in basic sanitation systems in urban slums in areas with strong governmental institutions that will sustain the network won't reduce disease, improve quality of life and increase worker productivity? Does anyone really think that US cold war aid to Zaire was primarily designed to promote economic growth, or that World Bank support for transport projects in Afghanistan is not in the slightest motivated by attempts to increase incomes?
So is anyone really that surprised that when you stick imperfect measures of aid flows into the right hand side of a regression along with a varying bunch of weak proxies for policy and institutional environments, poor indicators of educational and physical capital investment, and perhaps a few random variables covering geography and climate, then stick half-made-up and heavily manipulated figures for GDP per capita on the right hand side, then vary the country sample and the time period under investigation, and then vary the regression technique, your results sometimes suggest that aid can be harmful, and sometimes that it can be useful and frequently that we can't say anything with 95 percent confidence? If the results didn't suggest that, it would be the sign of a fix-up.
We all know that the answer to the question 'does aid work?' is 'that depends'. And the answer to the question 'does aid work better in good policy environments?' is 'that depends', as well. {emph added}
What William Easterly has called the "elusive quest for growth" is an understandable instinct for those committed to poverty reduction and economic development. Even with continuing increases in inequality in many developing countries, the rising tide does indeed lift the vast majority of boats, and few boats are going to be lifted in the absence of a rising tide.
Yet, given the vast number of variables and the imponderable causal links and feedbacks in even the smallest and least developed of economic systems, I've always found it misguided at best to focus on macroeconomic indicators as guides for assessing if, when, and how assistance is or is not "effective." Granted I have a strong personal bias for bottom-up, patient approaches, but why should we depend on macroeconomists to "validate" development policies? We have an ever-growing mountain of documented cases of development assistance -- with outcomes that are good, bad or ambiguous -- which can serve as inspiration and "lessons learned" for locals and assistance providers alike. Continuing to manipulate macrodata isn't going to give us reliable answers to "it depends on what?" As Adam Smithee notes:
It appears, sadly, that reality is ill-suited to the cross-country regression technique --so let's move on.
The first afoe European weblog awards