ADVANCEMENT OF THE SCIENCE
TABLE 1
Primary Data Sources Used
Data Type
Source/Platform
Access Link
Notes
PM 2.5 annual concentration (county level)
U.S. EPA Air Quality System https://www.epa.gov/air-trends/air-quality-cities-and- counties Direct link to the raw data: https://bit.ly/4wBZ20c
Main source for pollution values, some counties missing records
PM 2.5 monitor trends and pollutant listings Pollution ranking (initial county screening) Crime rate (per 100,000 population) Crime rate (historical time series)
U.S. EPA AirData Tools
https://www.epa.gov/outdoor-air-quality-data
Used to identify highest-pollution counties Used to identify top polluted counties, state mapping Primary official crime dataset, missing years encountered Used to supplement missing FBI records Used to cross-check values for incomplete series Used for the historical poverty rate per state Used for the historical unemployment rate per state
AirVisual, World AQI, U.S. EPA reports FBI Crime Data Explorer
U.S. EPA regional webpages, https://www.iqair.com/us
https://cde.ucr.cjis.gov
Macrotrends Crime Statistics
https://www.macrotrends.net/global-metrics/countries/ ranking/crime-rate-statistics
Additional crime verification USAFacts
https://usafacts.org
Poverty rate
U.S. Census Bureau
https://www.census.gov/data/tables/time-series/demo/ income-poverty/historical-poverty-people.html https://fred.stlouisfed.org/release/tables?rid=116&eid= 840687#snid=840690
Unemployment rate per state
Federal Reserve Bank of St. Louis
Note. AQI = Air Quality Index; FBI = Federal Bureau of Investigation; U.S. EPA = U.S. Environmental Protection Agency.
to a specific U.S. state in a given year. The ana- lytical sample covers 10 states in the nation observed annually during 2000–2022, yield- ing approximately 230 state–year observa- tions, which were subject to minor data avail- ability constraints. Unlike earlier exploratory specifications that relied on annual cross-state averages, the final analysis retains the full panel struc- ture of the data, preserving both within-state variation over time and cross-state heteroge- neity. This approach substantially increases statistical power and reduces the risk that estimated relationships are driven by shared national trends rather than underlying state- level dynamics. Variables and Measurement The dependent variable is the annual crime rate, measured at the state level and expressed per 100,000 population. The key independent variable of interest is the annual mean PM 2.5 concentration (µg/m³). To account for socioeconomic conditions that could confound the relationship between air pollution and crime, the analysis includes state-level poverty rates and unemployment rates as control variables. All variables are
TABLE 2
List of the Top 10 Most Polluted Counties and Their Respective Cities and States
Rank # Most Polluted County
City
State
1
Kern
Bakersfield
California
2
Lane
Eugene
Oregon
3
Oklahoma
Oklahoma City
Oklahoma
4
Richmond
Augusta
Georgia
5
Cook
Chicago
Illinois
6
Harris
Houston
Texas
7
Allegheny
Pittsburgh
Pennsylvania
8
Cuyahoga
Cleveland
Ohio
9
Clark
Las Vegas
Nevada
10
Maricopa
Phoenix
Arizona
Crime s,t = α + β 1 PM 2.5s,t + β 2 Poverty s,t + β 3 Unemployment s,t + ε s,t where s indexes states and t indexes years.
measured on an annual basis and matched consistently across states and years.
Empirical Specification The baseline pooled regression model is spec- ified as:
To address the potential for spurious cor- relations driven by common time trends,
20
Volume 89 • Number 3
Powered by FlippingBook