Saúde e desenvolvimento: três ensaios sobre condições iniciais

Abstract

This thesis is composed of three essays that analyze how the health shocks in the initial periods of life affect the accumulation of human capital. The first article, "Long-term effects of malaria on educational outcomes in the Brazilian Amazon region" applies the differences-in-differences estimator to explore the near-exogenous variation in malaria risk in space and time between cohorts located in municipalities where there was a change in malaria risk classification. The databases used are Prova Brasil (Inep- 2007 and 2011) and Sismal (Malaria Control Information System). The hypothesis is that being born in an area of high or medium risk of malaria implies an imminent exposure to the disease. The results show that this exposure at birth may be negatively correlated with proficiency in Portuguese and Mathematics. In the second essay, "The effect of the expansion of health services on the nutritional and health status of the children of Minas Gerais: a case study on the ‘Viva Vida’ Program", two analyzes were carried out. The method of differences in differences is used to verify if the expansion of the offer of assistance procedures, provided by the inauguration of the ‘Centros Viva Vida’ (CVVRS) in the state of Minas Gerais, contributed to the generation of a health effect on the population. The hypothesis is that CVVRS contributes directly to the reduction of the Infant Mortality Rate and the Maternal Death Ratio as it increases the supply of care procedures in secondary care, thus minimizing deaths due to preventable causes. The results show that the regions covered by the CVVRS showed a reduction in the indicators in relation to the municipalities that are not covered. The second analysis assesses the impact of the ‘Viva Vida’ Program on the health of children aged six to 48 months. The hypothesis to be tested is whether the actions implemented in the three levels of health care were able to provide health gains in the cohorts exposed to the program, evaluated by the anthropometric measures: Weight by Age, Height by Age, Weight by Height and Mass Index Corporal. These indicators were extracted from the Pesquisa de Orçamento Familiar (POF - 2002/2008). We used the combination of two estimation methods: propensity score matching and differences in differences. The results indicate that there was no reduction in the proportion of undernourished children and in severe malnutrition among the cohorts exposed to the program in relation to the non-exposed cohorts. Finally, the essay "The effect of the contribution of the mother to the family budget on the nutritional outcomes of the children" investigates the extent to which the contribution of women to the family budget can affect the health outcomes of children under 60 months. It is assumed that women who do not contribute financially at home have only access to income generated by other members. While among those who contribute financially to the family budget, control over resources is directly related to the contribution share. Child health is measured by the probability of presenting malnutrition according to the nutritional indicators of Weight by Age, Weight by Height, Height by Age and Body Mass Index. Data were collected from the Pesquisa de Orçamento Familiar (POF - 2002/2008). The method adopted is logistic regression. The results indicate that the probability of malnutrition is associated with the percentage of the mother's contribution in the family budget, the age group of the child and the type of nutritional indicator observed.