Panel Paper: Racial and Ethnic Disparities in Air Pollution: A Study of Excess Emissions in Texas

Thursday, November 8, 2018
Johnson - Mezz Level (Marriott Wardman Park)

*Names in bold indicate Presenter

Zhengyan (Ian) Li1, David Konisky2 and Nikos Zirogiannis1, (1)Indiana University, (2)Georgetown University


In recent years, environmental justice has been a focal point of attention of academia, policymakers, and most importantly the affected communities. Growing evidence suggests that many communities of color and low-income are disproportionately exposed to a variety of sources of air pollution. This paper contributes to this literature through an analysis of an under-studied type of air pollution: excess emissions. Excess or upset emissions are releases that occur during startups, shutdowns, or malfunctions. This type of pollution has historically been under-regulated in the United States, even though excess emission events, across many industrial sectors, are frequent in number and large in magnitude.

In this paper, we examine whether facilities located in communities with higher percentages of African-American, Hispanic, and poor populations tend to have higher levels of excess emissions. To do so, we compile a panel dataset of excess emissions and demographics for all operating industrial facilities in Texas from 2000 to 2010. The excess emissions information and facility characteristics are from the Texas Commission on Environmental Quality. Demographics for 1 mile buffer and 3 miles buffer around facilities are calculated with areal apportionment method based on tract level information from 2000 Census and 2010 Census.

We employ a two-part model for analysis, in which the first part predicts whether or not excess emissions occur and the second part models the amount of excess emissions conditional on facilities having nonzero excess emissions. Both models use the same set of demographic variables. We then combine the marginal effects from the two parts to calculate a combined effect and use a bootstrap method to calculate the standard error for the combined effect.

Our preliminary analysis shows that facilities in communities with higher percentage of Black and higher percentage of Hispanic consistently have more excess emissions across pollutant categories and buffer choices. However, contrary to the literature, we find that facilities located in communities with higher median household income tend to have more excess emissions.