Estimating the impact of pollution on wages and housing prices using satellite imagery

ABSTRACT

Air pollution is a global concern. However, research and policy addressing the problem remains hampered by a lack of ground-level air quality data. We estimate the value residents of Mexico City place on clean air using aerosol optical depth (AOD) from satellite imagery as a proxy for air pollution. We employ a unique dataset that combines actual household-level data with AOD and rainfall within a hedonic model focusing on people’s decisions regarding where to work and live in a developing nation. We address endogeneity and simultaneity concerns by using a 3SLS-IV estimation. Results suggest that residents place significant value on clean air, with a 10% increase in AOD leading to a 1.3% decrease in home prices and a 2.3% increase in wages.

KEYWORDS:

• Aerosol Optical Depth (AOD)
• air quality valuation
• hedonics
• Mexico City

JEL CLASSIFICATION:

• Q53
• R21
• O10

Disclosure statement

We declare there is no potential conflict of interest for this project.

Data Availability

A portion of our main dataset comprises of publicly available data from the National Water Commission (CONAGUA, https://www.gob.mx/conagua) of Mexico and the Level-1 and Atmosphere Archive & Distribution System (LAADS) Distributed Active Archive Center (DAAC), located in the Goddard Space Flight Center in Greenbelt, Maryland (https://ladsweb.nascom.nasa.gov/). The other component of dataset is not publicly available as it contains personal information that violates individual privacy. These data could be available on request from the authors removed of all personal identifying information.

Notes

¹ See Fontenla, Goodwin, and Gonzalez (2019) for greater detail of the empirical estimation and data description.

² We use a 2-SLS specification to estimate the Kleibergen-Paap statistic, where we get an F-value of 282.57, indicating rainfall is a relevant instrument. Further, to be able to perform the overidentification test, we add rainfall lagged by 1 month as a second instrument. The Hansen J statistic supports the validity of our instruments. Results available upon request.

³ OLS, 2-SLS, and results with additional controls available upon request.

Additional information

Funding

We declare there was no funding for this project.