Today I attended a seminar where Elsa Artadi presented her paper on: “Going into Labor: Earnings vs. Infant Survival in Rural Africa.” Artadi asked the question ‘why do families not optimize childbearing to coincide with months of minimal infant mortality?’ Artadi demonstrated that infant mortality rates vary significantly from month to month in Sub-Saharan Africa due to 1) the variation in disease incidence mostly from changes in rainfall and 2) agricultural cycles. Often mothers are faced with a trade-off between earning more income by working during the harvest season versus having a child during the low infant mortality months which may coincide with the harvest season. This is a serious tradeoff since giving birth in a low infant mortality month may not be optimal if the loss of income jeopardizes the health and well being of the entire family. For some countries, the low infant mortality months coincide with non-harvest months. Thus, these countries do not face this tradeoff and Artadi expects these individuals to concentrate births in the low infant mortality months.
Artadi constructs a ‘probability of survival’ variable for each country and month based on 1) whether that month was part of the rainy season, 2) whether that month was a high risk month for famine, 3) a fixed effects dummy for each mother, and 4) demographic characteristics. After calculating the fitted values of the expected survival rate (E[Surv]) for each month in each country, she creates another variable which measure the loss of infant health due to being born in a poor month (LossE[Surv]_m,c=max(Surv)_c-E(Surv)m,c). She then constructs a ‘Tradeoff_c’ variable which measures the difference between the expected survival rate during the high labor demand season (harvest) and the expected survival rate during the low labor demand season (non-harvest). Artadi runs a regession of ‘LossE[Surv]_m,c’ on ‘Tradeoff_c’, and demographic variables to see if citizens in countries whose low infant mortality months coincide with the harvest season choose poor survival months more frequently.
Artadi presents persuasive evidence that the sacrifice of potential income is the major reason why births would be concentrated in high infant mortality months. Although she had some technical problems (incorrect standard errors), her specifications of the model to check for robustness were convincing.
Significance of the Paper
While this paper provides a good description of the tradeoffs Sub-Saharan African women face in childbirth decisions, it does not offer any policy options for improving the welfare of these societies. Changing agricultural cycles is infeasible; most families already have some access to family planning (whether modern contraceptives or more traditional methods). In my opinion, the best means to decrease infant mortality in developing countries is to improve overall GDP/capita (easier said than done). Increasing incomes above subsistence levels will: 1) allow families to save so they will not go hungry if they decide to have a baby during harvest time 2) increase the available amount of money in a country to be spent on education and literacy which will help women become more knowledgeable about their health and 3) facilitate a demographic transition where more resources are devoted to each child. As an example India’s GDP has been growing steadily since 2000 and its infant mortality rate has decreased from 64.9 deaths/1000 live births in 2000 to 56.3 deaths/1000 live births in 2005. Further, the most famous micro-credit organization, Grameen Bank, claims that its small loans have lifted 50% of its members above the poverty line. In a report by H. I. Latifee in 2003, infant mortality decreased by 34% among Grameen Bank members. While part of this effect may be due to the educational programs Grameen offers, much is certainly attributable to the increase in income.