Vulnerabilidade hídrica e saúde neonatal: o impacto do projeto de transposição do rio São Francisco sobre peso ao nascer na região de Campina Grande

Abstract

This study analyzes the relationship between the expansion of access to potable water promoted by the São Francisco River Integration Project (PISF) and neonatal health outcomes, with a focus on birth weight. The analysis concentrates on municipalities within the Paraíba River Basin, with particular emphasis on the Campina Grande region, which experienced a prolonged period of water rationing between 2014 and 2017, coming close to a collapse in water supply. Microdata from the Live Birth Information System (SINASC) are used for the period from 2010 to 2021, allowing for the examination of birth weight dynamics before, during, and after the arrival of water from the PISF. The definition of treatment groups and the empirical strategy is based on the use of geospatial datasets from the National Water Agency (ANA), including shapefiles of rivers, basins, reservoirs, and project axes, as well as

official information on the release of PISF waters obtained through federal freedom-of- information portals. The empirical strategy employs a difference-in-differences approach with

synthetic control, adopted due to the difficulty of identifying observable control groups with pre-treatment trajectories comparable to those of the treated municipalities. The results indicate that there is no effect of the arrival of PISF waters on mean birth weight in any of the treatment groups. However, the results for median birth weight indicate negative and statistically significant effects in the first two treatment groups, with a stronger impact in Campina Grande and neighboring municipalities. At the 90th percentile, negative effects significant at the 1% level are observed only for this group, with reductions close to −48 grams. In placebo-controlled inference robustness tests, the results for Campina Grande and neighboring municipalities remain statistically significant for both the median and the 90th percentile. By contrast, for the treatment group composed of municipalities crossed by the Paraíba and Monteiro rivers, the coefficient for mean birth weight becomes statistically significant at the 10% level under placebo inference, indicating a reduction of approximately 23 grams.