By Birzhan Batkeyev David R. DeRemer

We find that air pollution increases crime in a city that ranks in the worst two percentiles worldwide for dirty winter air. Our identification strategy employs distinct geographic features of Almaty, Kazakhstan: cleaner mountain winds and frequent temperature inversions. Using these variables to instrument for PM2.5 air pollution, we estimate a PM2.5 elasticity of the expected crime rate more than four times as large as similar estimates from cleaner cities. Among crime types, we estimate statistically significant effects of air pollution on property crime, and we find no evidence of an effect on violent crime. These results are consistent with theory that air pollution induces higher discounting rather than aggression. We extend this theory and find that whether air pollution has distinct effects on crimes of varying severity depends on whether the population is more heterogenous in the outside option or in the discount factor. Using microdata on crime severity, we find statistically significant increases in both major and minor crime rates from air pollution, and we fail to reject common PM2.5 elasticities of minor and major crime rates. The greater scale of major crimes implies that they contribute more to the total crime rate increase from air pollution. 

Accepted manuscript for the Journal of Economic Behavior & Organization 2023.  

By Tatiana Zárate-Barrera

This paper estimates the non-monotonic effects of air pollution on criminal activity in a developing country setting and provides empirical evidence on the potential behavioral responses mediating this relationship. To do so, I combine daily administrative data on crime, air pollution, and sentiment polarity from millions of social media posts in Mexico City between January 2017 and March 2020. The identification strategy relies on highly dimensional fixed-effect models, non-parametric estimations of dose-response functions, and an instrumental variable approach that employs wind speed and wind direction as instruments for air pollution. My results suggest a causal and inverted U-shape relationship between air pollution and crime. Specifically, there is an inflection point after which marginal increases in air pollution negatively affect criminal activity. Exploring the mechanisms behind this relationship, I find that air pollution has the power to shape people’s emotional states and mobility patterns. These results provide important insights for developing countries where pollution levels are dangerously high, and crime is still one of the most pressing issues. In particular, under certain circumstances, environmental regulation tailored to reduce air pollution must consider the presence of behavioral responses and these non-linear interactions with criminal activity in their design. 

Unpublished paper, 2021. 48p.