UC Davis

My dissertation focuses on health outcomes for Chile during the 20th century.

My first chapter examines the effect of sulfonamide drugs on mortality and later-life outcomes. In particular, it explores the impact of sulfa drugs on reducing mortality and the long-term effects of being born in a better disease environment during the year of birth. I focus on Chile during the first half of the twentieth century.

The chapter starts by studying the contribution of sulfa drugs to the decline of mortality in Chile. Using new yearly data by province, including the cause of deaths, I estimate the impact of sulfa drugs on mortality, using a difference in difference approach for the period 1930-1950. For several infectious diseases, including pneumonia, sulfa drugs represented the first effective treatment. To see how sulfa drugs helped decrease Chile's mortality, I compare mortality from diseases treatable with sulfa drugs (maternal mortality, pneumonia, and meningitis) versus those unaffected by sulfa drugs (tuberculosis) before and after the adoption of sulfa drugs. Sulfa drugs are an excellent way of testing this because their adoption was fast because of their low cost.

I find that the introduction of sulfa drugs caused a considerable decline in pneumonia, meningitis, and maternal mortality. Specifically, sulfa drugs resulted in a drop of 10-28\% in maternal mortality. They also led to a 25-50\% decline in pneumonia mortality and a 10-40\% in meningitis mortality.

In the second part of this chapter, I use the introduction of sulfonamides drugs to identify the causal impact of exposure to pneumonia in infancy on later-life outcomes in Chile. There is a consensus that early life shocks can have persistent effects on later life (Barker (1992), Almond (2006), Bhalotra and Venkataramani (2015), Venkataramani (2012), Cutler et al. (2010), Bleakley (2010), Barreca (2010) and Lucas (2010)). My identification strategy exploits the introduction of sulfa drugs to identify the causal impact of exposure to pneumonia during infancy on later life outcomes. The idea is that being exposed to a better disease environment during childhood has short-term and long-term benefits. In the short term, mortality declines, but in addition, there are also long-term benefits associated with a healthier overall population because of this low endemicity.

My results show that exposure to sulfa-drugs, and thereby less exposure to pneumonia in the year of birth, led to a statistically significant improvement in education and employment for men. For years of schooling, a decrease of one standard deviation in pneumonia exposure (mortality) is estimated to have increased in 0.5 years of schooling for men (my results are not significant for women). The same effect is observed for my other educational variables. For employment, men born in an environment with a lower incidence of pneumonia are 2.8 percentage points more likely to be employed. I do not find significant results for the disability variable or mental disability.

My second chapter focuses on the impacts on the decline in mortality on fertility, labor markets, and marriage outcomes, using as a natural experiment the introduction of sulfa drugs in Chile in 1938.

Literature studying the effects of mortality declines on fertility is mixed (Soares (2005) argues that mortality decline will generate fertility decline, however Galor (2011) show more inconclusive results).

In theory, declines in infant and child mortality can have an ambiguous effect on fertility. For example, reductions in infant mortality may have a negative impact on fertility if parents have a preference for the target number of live births. If fewer children die, the probability of survival increases, and less precautionary childbearing is needed to get to the target number of children. But at the same time, a decline in infant mortality can positively affect fertility by reducing the price of a child's quality.

The effects of a decline in maternal mortality will also have ambiguous results in theory. For example, if maternal mortality declines, women's risk of dying will be lower, and this may increase fertility because the cost of having children is lower. This can also be interpreted as a decline in the price of child quantity. However, fertility can decline given that a lower risk of dying also increases expected life expectancy, so the benefits of getting educated increased. This is because the incentives of getting educated increase as the investment return also increases.

Using the 1960 and 1970 Censuses of Chile, I identify women of reproductive age during the period around the introduction of sulfa drugs and examine those cohorts of women when they have completed their fertility. I estimate models for the total number of children, distinguishing between the extensive and intensive margins of fertility. I also use a similar estimation strategy to analyze the impacts of mortality decline on labor and marriage market outcomes.

I show that child mortality decline, measured as pneumonia decline because of sulfa drugs availability, can decrease fertility by stimulating labor force participation. At the same time, a drop in maternal mortality also decreases fertility and increases the likelihood of remain childless. These results imply that the opportunity cost aspect because of longer life expectancy and higher returns of education and employment is more important than reducing the risk of dying during childbearing.

In particular, for intensive margin, an interquartile decline in pneumonia mortality (a movement from the 75th percentile to the 25th percentile), evaluated at the average number of reproductive years of exposure to sulfa drugs, led to 1.01 fewer total births for the average woman. The decline in maternal mortality led to 0.98 fewer births. While, for the extensive margin, a reduction in pneumonia mortality led to a 0.11 percentage point increase in the probability of being childless, while the decrease in maternal mortality a 0.26 percentage point increase.

Also, I show that declines in mortality from pneumonia increases labor force participation and employment status.

In terms of the marriage market, a decline in pneumonia and maternal mortality reduces the likelihood of a woman ever having married, consistent with the higher probability of being childless.

Finally, my last chapter studies the determinants of infant and maternal mortality for the period 1930 to 1960. Infant mortality rate (IMR), defined as infant deaths under one year old over 1,000 live births, is one of the most critical health outcomes. Moreover, it is not only important as a health outcome but as a social and economic development indicator.

Every country saw declines in infant mortality during the 20th century; however, the magnitude of the reductions was quite different. While there are still countries, like India, with infant mortality rates over 40 or countries in Africa over 50, there are some that have more successful histories. Chile is a clear example of this, the infant mortality rate in 1940 was one of the highest in the world, with over 200 babies dying before their first year of life. However, since then, IMR has declined significantly, Chile's IMR nowadays is similar to those of the developed countries (and lower than any other Latin American country).

Using panel data and an instrumental variable approach, with yearly hand-collected data from the Demographic Yearbooks of the National Institute of Statistics for infant mortality, maternal mortality, births, illegitimate births, number of hospitals, number of deliveries occurring in a hospital, number of deliveries with the assistance of a midwife/physician, for the period 1933-1960. I estimate a fixed-effect model of the impact of access to health care on infant mortality rate.

Results show that being born in a hospital reduces neonatal mortality. However, it doesn't have the same strong effects reducing total infant mortality, and it shows no effect on maternal mortality. The same holds for more presence of doctors or midwives. Also, my results are stronger for urban than rural areas.