Three essays in environmental health

This dissertation examines early childhood malnutrition, focusing on the role played by environmental shocks in determining health outcomes of children exposed to them. In the light of rapid climatic changes affecting human capital, Chapter 1 is one of the first to explore the prenatal channel of exposure to the December 2004 Indian Ocean tsunami, to study its impact on the long-run health status of very young children in India. It exploits variation from the region and timing of birth for Indian children using information from the Demographic and Health Surveys. This chapter finds evidence of a negative impact of an in utero exposure to the shock on important long term health indicators like birth weight and height-for-age, and others like weight-for-age, and child size at birth. This finding persists across various specifications, alternative control groups, and other robustness checks. I also examine the variation of this effect across birth trimesters, to find that for the more critical birth weight outcome, the first 2 trimesters are seen to be decisive in determining the extent of shock-based damage. It also finds that household wealth and maternal resilience, exhibited by mothers' health, education and employment, play an important role in alleviating the negative effect of the shock. This finding has education policy and labor market consequences, since it demonstrates how background can play a critical role in defining the extent of damage inflicted early in life by sudden shocks. Further, this chapter uses a unique quantile specification to find that children on the lower end of the height distribution, who are less resilient, suffer maximum damage. In addition, there is evidence of a son preferential gender bias in the shock impact in this lower tail. This demonstrates the significance of investigating gender bias in the effect of such shocks across the entire health outcome distribution, rather than only focussing on the mean. Chapter 2 uses Ghanaian household survey data to examine the link between environmental factors and the long-term health status of children in Western Ghana by exploiting the variation induced by a mining-based cyanide spill into a major water body, in October 2001. Information in the survey about the region and cohort of birth gives the primary source of identification. To get an improved spatial identification of exposed children, I link village level GPS data to the survey data. A rich set of controls are included to address omitted variable bias, by accounting for variables that can potentially confound the impact estimate if left unidentified. I also examine quantile treatment effects using a flexible specification not used before in this literature. This allows for uniquely incorporating the GPS information by smoothing on location. Findings reveal that after controlling for birth region and cohort, household, maternal and environmental factors, children born during the shock in the WR are negatively affected with reduced height, and that this negative effect persists through all the baseline and alternative specifications. Younger children are more affected than older ones. Finally, accounting for GPS information in the quantile specification yields a more negative coefficient for children on the lower end of the height distribution. Finally, Chapter 3 investigates the role played by a novel, structured variable selection mechanism not used before in child health literature, to identify childhood malnutrition risk factors. This method implements structured variable selection. By preserving the underlying structure of the input factors, it extracts more information from the potential set of candidate risk factors at a lower cost. This chapter gives evidence of a better prediction error using the structured selection procedure, the composite absolute penalty (iCAP), relative to other rival procedures for risk factor evaluation using the 2008 Ghanaian DHS. Using an experimental evaluation, it also shows improved model error and selection performance for the iCAP for models under misspecification. Finally, quantile analysis reveals differential impact of risk factors on children at different levels of malnutrition, improving our understanding of how one can mitigate moderate versus severe malnutrition.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Economics

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2012

Author dc:creator

• Vachharajani, Vidisha

Contributors dc:contributor

• Bera, Anil K.
• Koenker, Roger W.
• Lubotsky, Darren H.
• Baylis, Katherine R.
• Akresh, Richard S.

• Child health
• Environment
• Additive models
• Quantile regression.

Statement dc:rights

• Copyright 2012 Vidisha Vachharajani

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/30894

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/30894