Fire in the fields, crime in the air

While the health and productivity costs of exposure to air pollution are well-known, the resulting impairment of cognitive function and increase in physiological stress can also elevate aggression in human beings. Analysing Indian data, this article finds that short-term pollution exposure in regions downwind of major rice-growing districts during the rice harvest season, leads to an increase in violent interpersonal crimes, public order offences, and crimes against women.

Air pollution persistently reaches dangerously high levels in India, posing severe risks to public health. Beyond its known health and productivity costs, evidence indicates that air pollution can impair cognitive function and increase physiological stress, which can elevate aggression in human beings (Murphy et al. 2013). In recent research, we explore the potential link between rising air pollution and crime in India (Amale and Negi 2025).

In 2022, the National Crime Records Bureau (NCRB) reported over 52 lakh cognisable crimes, including approximately 35.6 lakh IPC (Indian Penal Code) crimes and 22.6 lakh Special and Local Laws crimes. Alarmingly, crimes against women reached 4.45 lakh, representing a 4% increase over 2021. In the same year, the ISRO (Indian Space Research Organisation) Protocol for monitoring paddy crop residue burning documented 53,792 burning events in Punjab, Haryana, Delhi and the NCR (National Capital Region) districts of Uttar Pradesh and Rajasthan, which are India’s leading rice-producing regions (Press Information Bureau, 2022).

The concurrent rise in crop residue burning and crimes highlight the larger social costs of intensive agriculture, which can exacerbate social instability and disproportionately affect vulnerable populations, particularly women and children in low-income communities. Since the late 1960s, government subsidies and price support policies have supported excessive specialisation in rice and wheat cultivation in India (Liu et al. 2021). The reality now is that due to excessive specialisation and rice-wheat cultivation system, farmers operate under tight post-harvest time constraints and often resort to burning leftover rice residues to prepare the fields for the winter wheat crop. This widespread practice generates sharp seasonal spikes in particulate matter (PM 2.5), producing some of the most severe air quality episodes in Northern India (Cusworth et al. 2018).

Our study: Methods and findings

We use high-resolution satellite data on wind directions, PM 2.5, and district-level crime records data to estimate the impact of short-term air pollution exposure on criminal activity. Establishing a direct connection between pollution from crop residue burning and crime is challenging because winds can carry smoke from burning fields over long distances. This spatial movement of pollution makes it difficult to identify whether changes in crime are driven by local factors or by pollution exposure from distant areas. We exploit changes in wind patterns that carry smoke from burning rice fields into neighbouring districts. These downwind areas may not practice residue burning themselves but still experience pollution exposure, allowing us to isolate the effects of pollution exposure on criminal activity.

Figure 1 shows that the wind direction changes exogenously over the years, carrying pollution from major rice producing districts to different regions every year. Based on these patterns, we define upwind districts as those where average Kharif (June to October) rice production from 2001 to 2019 falls in the top 25%, making them major contributors to crop residue burning. Pollution transported from these upwind districts provides exogenous increases in air pollution levels in downwind districts, allowing us to study the causal effect of residue burning related air pollution shocks on crime.

Figure 1. Wind direction in October

Notes: The arrows show the direction of wind from top Kharif (June to October) rice-growing districts in India, during 2001-2020. Rice producing districts are defined as those having average Kharif (June to October) rice production in the top 25% based on rice production between 2001 to 2019.

Figure 2 shows that the PM 2.5 levels in districts that are downwind of the rice-producing districts peak around 900-1,000 kilometres and then start to decline. In terms of magnitude, around the 900-kilometer distance band, downwind districts experience an increase of 4 µg/m² (or a 5% increase) in winter PM 2.5 levels. This pattern suggests that while pollution levels are higher in downwind districts, downwind pollution exposure declines with distance

Figure 2. Standardised PM 2.5 levels in downwind districts

Notes: (i) The figure displays estimated coefficients from separate regressions on the downwind ‘dummy’ for district sub-samples located between 100 and 2,900 kilometres from the top rice-producing districts. (ii) The dependent or outcome variable is standardised PM 2.5 levels. (iii) 95% confidence intervals are based on standard errors clustered by both downwind and top rice-producing districts. A confidence interval is a way of expressing uncertainty about estimated effects. A 95% confidence interval means that, if you were to repeat the experiment with new samples, 95% of the time the calculated confidence interval would contain the true effect.

Figure 3 shows the estimates of increase in crimes by different categories. We find that average additional downwind pollution (PM 2.5) exposure of 4 μg/m² leads to a 1.8% increase in overall crimes. In terms of numbers, this translates into approximately 24,000 to 26,000 additional crimes in 2020. This increase is driven primarily by violent interpersonal crimes, public disorder events such as arson and rioting, and most worryingly, crimes against women. We also observe a rise in property crimes, while economic offences, such as fraud and counterfeiting, remain unaffected. Back-of-the-envelope calculations suggest that pollution exposure during the rice harvest season generates an unaccounted social cost of approximately US$600 million.

Figure 3. Estimated increase in crimes in downwind districts due to pollution exposure

Notes: (i) The figure presents the estimated increase in crimes in downwind districts due to pollution exposure from residue burning in the top rice-producing districts. (ii) The 95% confidence intervals are shown. (iii) Violent interpersonal crimes include murder, assault, and kidnapping. Property crimes cover robbery, burglary, and theft. Economic crimes include breach of trust, cheating, and counterfeiting. Crimes against public order include rioting and arson. Finally, crimes against women cover rape, sexual harassment, dowry deaths, cruelty by the husband or relatives, and assault on women.

Mechanisms

What could explain this link between pollution exposure and heightened criminal activity? As evident from the medical literature, one of the plausible mechanisms is that pollution exposure increases aggression by impairing cognitive function, thus reducing impulse control. Our results are consistent with this mechanism. We observe an increase in violent crimes such as assaults on women, rape, murder attempts, and kidnappings. This is consistent with the aggression channels, as these crimes are largely aggression motivated. In contrast, we find no significant effect on crimes such as cheating, counterfeiting, and breach of trust, which are generally planned and opportunistic.

A second mechanism could be visibility. Pollution reduces visibility due to haze, weakening deterrence and supporting opportunistic crimes in public spaces. Reduced visibility can decrease the likelihood of detection as fewer people may be present to serve as potential witnesses. We observe that average visibility during the three winter months decreases by 3% in downwind districts at the 900-kilometer distance band (Figure 4)

Figure 4. Visibility in downwind districts

Notes: (i) The dependent variable or explanatory variable is average visibility (in meters) during October-December. (ii) Coefficients are from regressions on the downwind dummy across distance bands (100–2,900 km). (ii) Confidence intervals are based on standard errors clustered by both downwind districts and top rice-producing districts.

Third, pollution can also create economic distress and can further lead to incentives for property crimes. We do find a 1.7% decline in the gross domestic product (GDP) from the services sector in districts exposed to the downwind air pollution. This pattern is consistent with reduced labour market participation and earnings, driven either by deteriorating air quality or due to increased perceptions of public insecurity.

Finally, we find that districts with a higher density of police stations help reduce the rise in property crimes. This makes sense, since property crimes are usually planned, and a stronger police presence can act as a deterrent. In contrast, violent crimes, which may be triggered by the cognitive and emotional effects of pollution, are less responsive to increased law enforcement. This suggests that better policing alone may not prevent these offences.

Conclusion

Our findings reveal a significant and previously overlooked social cost of residue burning and air pollution in terms of heightened crime in India. A more concerning finding is that downwind pollution exposure also leads to an increase in violent crimes against women. This carries broader implications for understanding the constraints on women’s economic and social participation. While academic discussions on low female workforce participation often focus on factors like marriage, childcare, and gender norms, our findings suggest that rising pollution itself may discourage women from leaving their homes not only due to health concerns but also because it may contribute to a more unsafe and violent social environment.

Our results support the suggestion to shift the focus of environmental interventions from targeting the most polluted areas to prioritising the most vulnerable populations, who may suffer disproportionately from pollution exposure. The evidence also highlights the importance of integrating environmental quality into policy discussions on safety, gender equality, and women’s empowerment.