This paper studies how exposure to crime affects demand for policing using a unique setting where both crime and demand can be measured at the neighborhood level. Specifically, I use precinct level returns from ballot measures in San Francisco to provide the first causal evidence on how individuals’ support for pro-police policies responds to exposure to crime. Using variation in criminal activity across neighborhoods around election day, I find that each additional violent crime leads to an increase in support for police union-endorsed ballot positions ranging from 2.9 percentage points for homicides to 0.4 percentage points for lesser crimes. The effects are present during biennial congressional elections but not during municipal elections, suggesting the results are driven by lower-propensity voters. The effects are also largest in areas with high shares of white residents. 

We study how the arrival of new information on disaster risk produces long-run changes in the economic structure of affected communities. Leveraging plausibly exogenous variation in flood risk map updates, we use an event study approach to show that provision of more accurate flood risk information causes long-run decreases in property crime. We also demonstrate that communities which receive updated flood risk information experience long-run changes in population, house prices, median income, and composition of business establishments. We test several plausible mechanisms for the observed crime reduction, and we find these economic changes are most consistent with an eyes-on-the-street theory of crime. The effects of flood risk information are amplified when the updated information arrives after a physically destructive flood, suggesting a role for information salience. Together, these results show disaster risk information produces lasting changes in economic conditions.

Does exposure to climate-related natural disasters affect individuals' willingness to engage in climate activism? Prior studies show that exposure to high temperatures and natural disasters affects pro-climate preferences, but there is little evidence on whether it changes individuals' willingness to participate in costly, public actions to influence climate policy. We assemble a novel panel dataset of nearly 150,000 reported attendances at more than 5,000 pro-climate events in the United States. Using this panel and a stacked event study approach, we show exposure to climate change-related natural disasters such as large wildfires and hurricanes causes an increase in attendance at pro-climate events. We also show that attendance increases in distant areas that are not physically affected by a disaster but are socially connected to disaster areas. Furthermore, public exposure to large pro-climate events itself transfers through the social network, producing increases in activism at distant locations. 

Despite extensive evidence that extreme temperatures impose substantial economic costs, environmental and economic concerns are often viewed as conflicting. Using plausibly exogenous year-to-year temperature realizations, we find that extreme temperatures increase vote share for Democratic presidential candidates. These effects are larger in counties with greater economic exposure to temperature shocks—areas that also experience larger wage declines. We further show that the response operates through mobilization of Democratic-leaning voters rather than persuasion of Republican-leaning voters. Together, these findings suggest that voters respond to the economic costs of climate inaction, indicating that environmental and economic considerations can be complements.

Although waterfront property now commands a premium in modern cities, rivers and lakes near dense settlements were befouled with sewage for much of human history. This paper studies how the removal of human waste from surrounding waters transformed the urban environment using the engineering feat of the Chicago River reversal in 1900, which produced rapid improvements in surface water quality in a dense urban area. We find that rental prices were highly responsive to sewage exposure. The permanent reversal of the Chicago River increased listed rental prices by over 30% for properties nearest to the lake shore. This effect decreases with distance from the lake and is persistent over time. We also provide descriptive evidence that better water quality led to improved amenities and an influx of professional workers to residences near the lakeshore.

This study analyzes the effect of acute sleep disruptions on the high--stakes standardized test performance of high school students. I leverage a natural experiment arising from a 2007 policy change in the United States which moved the spring clock change into Daylight Saving Time (DST) immediately prior to statewide graduation exams in Ohio. Using a difference-in-difference estimation with a panel of nearly 900 school districts, I find the clock change lowered graduation exam passage rates between 1.9 and 2.5 percentage points. Results suggest DST is not costless for adolescent students and contribute to a growing body of evidence on the negative cognitive effects of sleep disruption.

Governments monitor air quality for regulatory purposes and, more recently, to provide information so individuals can act to lower their exposure to air pollution. Recent developments in low-cost technologies have also led to private adoption of air-quality monitors that produce publicly accessible air-quality readings. We study the adoption of these private air-quality monitors. We find that shocks to air pollution from wildfire result in substantial adoption. We also find that additional private monitors are concentrated in white, wealthy, and politically liberal neighborhoods. In contrast, there is no evidence that pollution shocks lead to higher adoption in neighborhoods with lower pre-existing access to monitors, higher long-run pollution, or those with more vulnerable populations. Private provision increases inequality in the availability of localized air-quality information.

We estimate the impact of piped water and sewers on property values in late-19th century Chicago. The cost of sewer construction depends sensitively on imperceptible variation in elevation, and such variation delays water and sewer service to part of the city. This delay provides quasi-random variation for causal estimates. We extrapolate ATE estimates from our natural experiment to the area treated with water and sewer service during 1874-1880 using a new estimator. Water and sewer access increases property values by more than a factor of two. This exceeds costs by about a factor of 60. 

How does exposure to risk shape individual preferences for an expanded state? I examine this question in the context of climate change-related risk. Using variation in California wildfire activity, I show neighborhoods experiencing large fires increase support by 0.8 percentage points for ballot initiatives which expand the size of government and by 2.4 percentage points for ballot initiatives endorsed by pro-environment interest groups. The effect is stronger in Republican areas and is not driven by shifts in voter registration or turnout, suggesting the mechanism acts through changes in individual preferences rather than compositional changes in the electorate.