Over at Apprehension, Douglas Bass offers another thought and an interesting graphic on the question of foreign aid and infant mortality rates (IMR).

The rows are nations/regions, the columns are years, and the cells are the IMR's for that nation/region and year. The cells are black if the Census Bureau doesn't have data for that nation/region and year, brown if the IMR is above 50, and green if the IMR is below 50. All of the countries that were green when measurement started have stayed green, and most of the countries that were brown, and have gone green, have stayed green.

Now what I plotted before, for Africa, was the level of IMR versus foreign aid. Douglas' post talks about the change in IMR. Given that he left on my comments page the data for his graph, I thought it rather simple to calculate the decline in IMR versus foreign aid. So, same graph except using the change in IMR over 1990-2000.

I've got a slight change in my data, as I added North Africa back into the dataset. I also have the problem that I only have aid data from my source for 1999-2003, and only the 2000 dataset looks complete. So I could be wrong in what I have here if the earlier aid pattern to Africa is different than it was in 2000. This is supposed to be a back-of-the-envelope exercise, not a detailed senior paper. (If any of my students are reading: You have a ready-made senior paper topic.) But here are two pictures, one for the level of IMR in 2000 (the previous graph had two fewer datapoints) and the second of the change in IMR over the 1990s versus foreign aid in 2000.

There doesn't appear to be much of a relationship. Depressingly, the data don't seem to show any change in IMR in the 1990s in Africa; you can count seven countries with noticeably hire IMR in 2000 than 1990.

This week's Economist has a review of the aid for Africa question as its headline piece, and it includes this intriguing reference to the Sharon Stone-mosquito bednet story.

Top of the list of quick wins are mosquito bednets, impregnated with insecticide. They cost less than $4 and cut the risk of infants dying by 14%, to 63%. The appeal is obvious and immediate. At the World Economic Forum in Davos this year, a speech on malaria by Benjamin Mkapa, Tanzania's president, prompted Sharon Stone, a Hollywood actress, to stand up, pledge $10,000 for bednets on the spot, and challenge her fellow audience members to do the same.

Sadly, this impulsive generosity will not be instantly gratified. Nets cost more to distribute than to make. Misguided policies can make matters worse. Nigeria, for
example, has on various occasions imposed tariffs of up to 40% on imported nets to protect its own netmakers. Demand for the insecticide, with which many Africans are unfamiliar, cannot be taken for granted (less than a fifth of nets are retreated regularly) nor can demand for the nets themselves. The Monitor, a Ugandan newspaper, reports that a government official last month warned villagers not to turn their nets into wedding gowns.

So where are Ms Stone's nets now? In fact, the Tanzanian government has a sensible policy of not giving bednets away. To do so might crowd out the commercial sellers of bednets, who distribute them more efficiently than the public sector�and can be relied on to keep selling them, provided they can make a profit, long after celebrity donors have lost interest. Instead, the government hands out vouchers to pregnant women at antenatal clinics, covering much of the cost of the nets in the market.

The dilemmas of distributing bednets illustrate some general problems of aid. Donors muster resources, but they fail to align the incentives of the people providing them or benefiting from them. The grand macro-solutions often neglect the nagging micro-foundations.

The Economist story goes on to tell of informal markets developing to distribute drugs; this may not be out and out corruption but simply the means by which clinics stay in business when they don't receive payments expected from the public sector. And getting incentives right is important for both the micro and macro levels, as this piece from the IMF's research chief Raghuram Rajan suggests. It's not just the amount that Africa gets but how it gets it that matters.

If the country�s government is thoroughly corrupt, then the status quo�no forgiveness and no additional aid�is best, for it gives the government no official resources to misuse and limits its ability to raise private sector funds. Aid in this case should be distributed directly to nongovernmental organizations. If the country has a reasonably committed government, look at the country�s primary need. When social sector projects top the list, then what matters is the extent of official sector net funding. Here, the first alternative�debt is not forgiven but official creditors lend more�is best. But if most projects are commercially viable, the second alternative�some relief but leaving enough outstanding official debt that foreign private investors lend responsibly�may be optimal. Finally, substantial debt forgiveness is prudent if the risk of financial distress really is a serious problem�an unlikely eventuality. But there must be an assurance that the country does not borrow up again from private creditors and game the system to get further debt relief.

The Africa question is daunting, as more and more advisors become frustrated and calloused with the lack of progress after pouring in so much and receiving so little. The story dates back as far as the end of WW2, when parts of Africa were seen as better prospects for economic growth than Japan. The Economist writers want us not to abandon all hope. We haven't, but we would like to be sure first that we do no harm.