Evidence on trade shocks in Brazil shows that, despite reducing household incomes, child health improved in hard-hit areas as parents, particularly mothers, had fewer labour market opportunities outside the home and therefore more time to invest in their children
One of the most striking trends over the last 40 years has been the dramatic decline in infant mortality across most of the developing world. Between 1990 and 2000 alone, the infant mortality rate dropped by about 21% in low-income regions. For some countries, such as Chad and Niger, this decline represents the first sustained reduction in infant mortality ever known.
This decline has coincided with the rise of globalisation, which has radically transformed the structure of labour markets and altered the decisions made by household. One manifestation of this phenomenon is the series of reforms conducted in the 1980s and 1990s that lowered import tariff barriers in many developing countries. Do economic shocks generated by changes in trade policy spur or slow infant mortality declines? In our recent research (Charris, Branco and Carrillo 2023), we examine this question by focusing on an unprecedented trade reform that caused sharp variation in economic conditions across local labour markets in Brazil.
In the early 1990s, the Brazilian government implemented a unilateral process of trade liberalisation that lowered import tariffs from an average of 30.5% in 1990 to 12.8% in 1995. The bulk of these declines occurred in sectors with high initial import tariffs, as seen in Figure 1. This tariff shock was unexpected, sudden, of considerable magnitude, and permanent. Our research design exploits variations in the intensity of the tariff cuts across industries and geographic differences in industrial structure, following the approach developed by Kovak (2013) and Dix-Carneiro and Kovak (2017).
The literature has documented that local labour markets disproportionately exposed to international trade experienced contractions in employment, lower incomes, and higher earnings inequality. Based on this observation, one could tentatively hypothesise that worse labour market conditions generated by increased import competition should increase infant mortality. However, literature in health economics suggests that parental time is a critical input in the child production function (Miller and Urdinola 2010). Worse labour market opportunities make it less costly to undertake health-improving behaviours that are time-intensive. If this opportunity cost effect is large enough, the overall impact on infant mortality could be negative. Whether trade-induced economic shocks increase or reduce infant mortality depends on the relative importance of these mechanisms and thus remains an empirical question.
Figure 1: Import tariff changes across industries.
Notes: This figure shows the incidence of import tariff cuts between 1990 and 1995 across industries (Charris, Branco and Carrillo, 2024).
Our empirical analysis uses data on infant mortality covering the 1985-2010 period at the municipality level. The choice of the period is determined on the back end by the availability of birth estimates from the population census, whose last round was in 2010. We also use individual-level information from Brazil’s 1980, 1991, 2000, and 2010 population censuses. This data provides demographic information, including gender, age, and race, as well as socioeconomic characteristics such as educational attainment, employment, occupation, industry categories, labour income, and non-labour income. We use census information to construct the regional measure of exposure to the trade reform and to explore changes in employment rate and mean household income.
Effects on infant mortality
We find a visually clear and statistically significant decline in the infant mortality rate in areas housing the industries with greater tariff cuts. Figure 2 shows that after the reform was introduced, there appears to be a decline in infant mortality rates that became highly significant in the post-liberalisation period, with a steadily growing effect pattern. The timing of these effects and their steadily increasing magnitude nearly mirror the labour market responses documented in Dix-Carneiro and Kovak (2017). However, the response of infant mortality is somewhat slower. Our estimates imply that a one standard deviation increase in the exposure to the trade shock is associated with a decline of 9 to 13% in the infant mortality rate.
Figure 2: Regional Tariff Reduction and Log-Changes in Infant Mortality Rates.
Source: Charris, Branco and Carrillo, 2024.
When we look at specific causes of death, infant mortality from infectious and respiratory causes decreased the most. A significant factor associated with this death category is water quality (Bhalotra et al. 2017), and trade-induced unemployment could have improved parental time to collect water of higher quality from distant sources. This is especially likely to be true in our setting, as a significant portion of households still have limited access to improved water quality from public systems and must thus spend considerable amounts of time to obtain pure water from alternative sources.
Potential mechanisms
We investigate the potential role of parental time in driving the observed decline in infant mortality. We begin by examining male and female employment. Our estimates imply that men appear to have adjusted to the trade-induced economic shock by transitioning into the self-employment sector, with no meaningful changes in their overall employment rates. In contrast, this adjustment response is not present among women. Women experienced a significant decline in the likelihood of working in exposed industries that is not compensated by an equivalent increase in the self-employment sector, suggesting that women have exited the labour market permanently and arguably allocated more time to home production activities.
When we examine changes in household income, we observe a statistically meaningful decline in harder hit locations. This suggests that male adjustment response has not been large enough to compensate for the income losses experienced by areas with greater exposure to the reform. These results are broadly consistent with the importance of parental time in the production of child quality. This interpretation is particularly plausible in the context of Brazil, where most preventive and primary health services are provided free of charge, but parents must allocate a considerable amount of time to obtain them (e.g. travelling to distant health facilities and long waiting times in primary care centers).
Consistent with this hypothesis, we find that exposure to the trade-induced economic shock is associated with an increase in the probability of preventive visits among women of childbearing ages, which include prenatal care visits, and an increase in the fraction of infants with growth-monitoring and well-care visits.
Key takeaways
We study the effects of changes in local economic conditions generated by a trade liberalisation reform on infant mortality in Brazil, using a rich dataset that covers more than 3000 municipalities over a horizon of 25 years. Our findings indicate that areas housing industries with greater tariff cuts experienced a decline in infant mortality, which is consistent with the hypothesis that worse labour market opportunities make it less costly to undertake health-improving behaviours that are time-intensive.
We show that while men have adjusted to the trade shock by transitioning to the self-employment sector, women have permanently exited the labour force and arguably allocated more time to home production activities. To some extent, this evidence is consistent with the importance of parental time for the household production of child health status. From a policy perspective, these findings suggest that policymakers promoting female labour force participation could mitigate some of the unintended consequences of parental work by introducing flexible programmes that facilitates access to primary and preventative health services even among parents who face significant time constraints. We find limited evidence in support of other potential mechanisms, including changes in the supply of health services, air pollution levels or consumption of harmful normal goods.
Editors' note: This column is published in collaboration with the International Economic Associations' Women in Leadership in Economics initiative, which aims to enhance the role of women in economics through research, building partnerships, and amplifying voices.
References
Bhalotra, S R, A Diaz-Cayeros, G Miller, A Miranda, and A S Venkataramani (2017), "Urban water disinfection and mortality decline in developing countries", Technical Report, National Bureau of Economic Research.
Charris, C, D Branco, and B Carrillo (2024), "Economic shocks and infant health: Evidence from a trade reform in Brazil", Journal of Development Economics, 166: 103-193.
Dix-Carneiro, R and B K Kovak (2017), "Trade liberalization and regional dynamics", American Economic Review, 107(10): 2908–46.
Kovak, B K (2013), "Regional effects of trade reform: What is the correct measure of liberalization?", American Economic Review, 103(5): 1960–76.
Miller, G and B P Urdinola (2010), "Cyclicality, mortality, and the value of time: The case of coffee price fluctuations and child survival in Colombia", Journal of Political Economy, 118(1): 113–155.