« Back to Results

Health Effects of Environmental Conditions

Paper Session

Sunday, Jan. 3, 2021 3:45 PM - 5:45 PM (EST)

Hosted By: Health Economics Research Organization
  • Chair: Janet Currie, Princeton University

The Impact of Clean Water on Infant Mortality: Evidence from China

Maoyong Fan
,
Ball State University
Guojun He
,
Hong Kong University of Science and Technology

Abstract

This study is the first national research on clean water and infant health in China. We have compiled what is to our knowledge the most comprehensive assessment of clean water access, infant mortality, and surface water quality in China ever assembled. By combining these three data sets, we are able to examine an accurate sample of infant mortality and clean drinking water at a level of quality rarely found in developing countries.
To address the endogeneity concerns of piped water coverage, we construct a novel instrumental variable based on the least-cost route from water sources to destinations using GIS network analysis. Because construction costs of water pipelines depend critically on the geographical characteristics between a mortality surveillance area and its nearby clean water sources (e.g. reservoirs), we can use the least-cost distance between them as the instrument for piped water coverage. Since the least-cost path is constructed purely based on cost considerations, it should be uncorrelated with demand-side determinants of infant mortality (such as income and social preferences).
We find that provision of piped water significantly decreases infant mortality with a 10 percentage-point increase in piped water reducing infant mortality by 16 percent. Combining surface water pollution and piped water coverage, we find that the effect of piped water provision is greater in regions with slightly polluted surface water rather than in regions with severely polluted water. This counter-intuitive result can be attributed to people’s avoidance behavior: people continue to use slightly polluted water while they avoid visually polluted water. We also show that the economic benefits of piped water in rural China are significantly greater than the estimated costs of its provision.

Coal Use, Air Pollution, and Student Performance

Valentina Duque
,
University of Sydney
Michael Gilraine
,
New York University

Abstract

Coal is both the primary source of global energy and of air pollution. This paper presents the first causal evidence of the impact of pollution due to coal power plant emissions on cognitive outcomes. Our approach combines rich longitudinal student data with a design leveraging year-to-year coal plant emissions, persistent wind patterns, and also plant closures. We find that every one million megawatt hours of coal- fired power production decreases student performance in schools within ten kilometers by 0.02 SD. Gas- fired plants exhibit no such relationship. Our analysis indicates that declining coal use has affected student performance and test score inequality substantially. We find that attending a school within 10-Km and `downwind' of a coal-red power plant lowers student performance in math and English by 0.08 SD and 0.04 SD, respectively. In contrast, we find no such relationship between natural gas power plants and student outcomes. We also extrapolate our estimates nationwide. Given that underprivileged students are more likely to attend schools near coal plants, our results indicate that the precipitous decline in coal use in the last decade has generated substantial improvements in student performance and equity.

Heat and Violent Behavior Among Prisoners in Mississippi

Nicholas Sanders
,
Cornell University
Anita Mukherjee
,
University of Wisconsin

Abstract

We examine how extreme heat drives incidents of violent behavior of prisoners in Mississippi. Our data cover all individuals in Mississippi prisons from 2004-2010, and include prisoner demographics, location, and recorded violent infractions by day. Our treatment uses daily, county-level temperature. We group average daily temperature in Fahrenheit (F) into degree-range bins. This allows us to test for non-linearities in the response to heat. Identification comes from unusually hot days across different parts of the state, after controlling for statewide month-by-year effects, day-of-week effects, and prison-specific time trends.
Relative to a day in which the average temperature falls in the 60-70 F range, a day with an average temperature in the 80-90 F range increases incidence of prison violence by approximately 23% on a baseline of 0.33 incidents per prison/day. It increases the probability of any violent incident by 1.7 percentage points, on a baseline of 16 percentage points. Both outcomes are robust to variations in time and region fixed effects.
Our results contribute to the literature on temperature and violence, with a unique setup that largely avoids two common confounders: avoidance behavior and reporting bias. Prison limits avoidance behavior; prisoners cannot leave the prison on hot days, and the prisons in our analysis do not provide air conditioning. Our setting also somewhat addresses reporting bias. In studies of temperature and violence using regular crime statistics, police may be less likely to patrol and potential witnesses may remain indoors, making crime reporting less likely on hot days. In our setting, prison workers must be in the prison monitoring prisoners regardless of temperature and potential witnesses (other prisoners) must remain as well.
Discussant(s)
Alan Barreca
,
University of California-Los Angeles
Michelle Marcus
,
Vanderbilt University
Jillian Carr
,
Purdue University
JEL Classifications
  • I1 - Health
  • Q5 - Environmental Economics