My Twitter followers are a very smart bunch, in general, and so I expected the wisdom of crowds to win out when I put this question to them on Monday.

Boy I was wrong. The true answer, as detailed in the first chapter of my friend Charles Kenny’s new book, is that Afghan life expectancy rose an astonishing 20 years between 2004 and 2010, from 42 years to 62 years. In terms of development aid effectiveness, it was one of the most astonishing wins of the 21st Century.


This public-health miracle didn’t even cost very much: about $4.50 per person per year, paid by the Afghan Ministry of Public Health with the help of USAID. That money vastly increased vaccination rates, saved 100,000 deaths per year of children under the age of 5, and improved health outcomes for adults too. If you’re one of the people looking for bang-for-the-buck case studies in development aid, Afghanistan from 2004 onwards is a great place to look, especially since the improvements managed to take hold in a notorious failed state.

Of course, it’s not easy to audit exactly where that money went, and in a country like Afghanistan, it’s easy to believe that some of it was trousered by corrupt officials. Which leads to the question at the heart of Kenny’s book: Is corruption, in and of itself, a reason not to do stuff? And can it be counterproductive to worry too much about corruption?

Certainly corruption can render certain actions useless, or worse. If you give a town money to buy much-needed electricity, but all that money is diverted to a local warlord who uses it to buy guns, then you have caused more harm than good.


On the other hand, if you’re clearly getting positive results from your actions, how much does it matter if you’ve marginally increased the amount of graft along the way? Suppose for instance that you give the same town money to buy electricity, and they do use it to buy electricity, but some of the electricity they buy has been stolen from the state-owned power company by employees illegally tapping the power lines. In that case, the good caused by the introduction of electricity would outweigh the harm caused by the corruption.

As Kenny says, it’s crucial not to use corruption as an excuse for “avoiding the moral responsibility to act”. If you ask Americans how much foreign aid ends up in the hands of corrupt officials, for instance, the median answer is 60%. That’s a great reason to cut foreign aid – or at least it would be, if it was anywhere near the truth, which it isn’t. What is true is that an enormous amount of foreign aid goes into internal compliance functions, and to US and other western companies which can prove their clean hands to the satisfaction of the western bureaucrats. Haiti is the most ignoble example: just 1% of the aid promised to the country in the wake of the 2010 earthquake went to the government. The rest trickled down through a network of NGOs, who ended up unhelpfully dominating the domestic economy as a result.

Which is where the charity angle comes in. Huge amounts of foreign aid end up flowing into the NGO sector, often to extremely efficient, effective, and well-run organizations. Still, every dollar of foreign aid which goes to NGOs is a dollar which doesn’t go directly to local governments. And that’s a problem.

After all, the Afghan government surely learned a lot in implementing the USAID program. If USAID had instead gone around the government, and instead asked say Partners In Health to do the same thing, then a lot of institutional knowledge about how to make the country healthier would have remained within PIH, rather than entering Afghanistan’s nascent bureaucracy.

It’s incredibly rare for individuals to target their philanthropic donations at national governments of any stripe. (The CDC Foundation is probably the most prominent exception to the rule, and it’s not exactly focused on raising small-dollar donations.) But when you give money to US or European charities working in developing-country public health, you are in a way saying that you don’t trust the local government to deliver results, and that you’d rather give your money to, well, a white savior.

In the Afghan case, the US Special Inspector General for Afghanistan Reconstruction ended up recommending that the health program be suspended, not because it found any evidence of corruption, but just because “financial management deficiencies” meant that they didn’t know whether, maybe, some corruption might be going on. Probably a sophisticated American NGO would not have had such deficiencies. But the Afghan program was working, and working very effectively. That’s ultimately what matters. When you add NGOs to the development mix, it’s important to remember what you’re subtracting at the same time.