The impact of the global financial crisis on infant mortality is a topic of great policy importance. However, estimates of the likely impacts of the crisis, cited by international institutions and in the popular press, differ wildly.
This blogpost summarizes the main conclusions from some of my own recent research on this topic, jointly with various colleagues.
These conclusions include:
1. The effect of negative growth on child health, including infant mortality, varies a great deal across countries. In developed countries, such as the US, infant mortality decreases when there are negative economic shocks. In low-income countries, including in India and countries in Sub-Saharan Africa, infant mortality increases in those periods (see Ferreira and Schady 2009 for a discussion ). The picture for middle income countries is mixed, but on balance, it is closer to that found in the US. Important exceptions are cases in which the economic contraction was very severe--maybe 15 percent or larger. That was the case in two crises in Peru, and in Indonesia in 1998 (see Paxson and Schady 2005  and Schady and Smitz 2009 ).
2. We estimate that the current downturn will result in somewhere between 30,000 and 50,000 excess deaths in Sub-Saharan Africa in 2009 (Friedman and Schady 2009 ). This is substantially lower than the numbers on Shanta's blog (700,000 excess deaths) and some of the very large (and I would argue implausible) numbers that are regularly cited in the press. The main reason for the discrepancy is the expected contraction in GDP that is used for the calculations. The numbers we estimate are based on the March 09 IMF projections, while the numbers on Shanta's blog are based on a "growth collapse", which amounts to a 22 percent decline in GDP. Such a collapse now seems unlikely: The June 09 projections from the World Bank's Prospects Group, for example, foresee that per capita consumption in Sub-Saharan Africa will increase by 0.74% in 2009, a modest decline in relation to the 2008 level of 1.87%. (These numbers are simple averages across countries, without population weights.) By next year, the growth rate is projected to be back to 1.95%.
3. Aggregate economic shocks generally have larger impacts on infant mortality among girls than boys -- see Baird, Friedman and Schady 2009 , and Friedman and Schady 2009 . We find this to be the case in every developing region: Sub-Saharan Africa, Latin America and the Caribbean, East Asia, South Asia, and the Middle East and North Africa. There are not enough countries with data in Eastern Europe and Central Asia to see whether this pattern holds there. This is a novel finding, and a very discouraging one. We also show that these differential mortality patterns by gender are very unlikely to have a "biological" explanation--girls are generally more sturdy than boys, and there are no significant or substantive changes in the boy-girl ratio at birth during crises. Rather, it seems, families make greater efforts to protect boys than girls in dire economic times.
The upshot of all of this is that there is no reason to think that the sky is falling down, in either poor or middle-income countries (at least in terms of infant mortality).
However, policies that protect girls would seem to be an important priority, especially in countries that are expected to face very deep economic contractions.