Author ORCID Identifier



Date of Graduation


Document Type


Degree Type



College of Business and Economics



Committee Chair

Brad R. Humphreys

Committee Member

Jane E. Ruseski

Committee Member

Daniel S. Grossman

Committee Member

Alexander Cardazzi


The first chapter analyzes how Premier League games in London impact demand for the city’s bike-sharing system. Bike-sharing systems affect mode choice for within-city transportation. Using bike rental data from Transport for London, this chapter exploits the plausibly-exogenous timing of Premier League games to identify how proximity to stadiums affects demand for London’s bike-sharing system. Results indicate games at Stamford Bridge (stadium) cause a 24.36% increase in the number of bike trips arriving within 0.5km of the stadium, with effects diminishing as distance increases. In contrast, games at London Stadium decrease nearby bike-sharing use and may actually crowd out would-be bike-share users closest to the stadium. Proximity to stadiums also changes cyclist behavior in other ways: depending on the stadium, cyclists arriving within 0.5km of a stadium on a game day travel 7.42-14.36% farther than the average day in the sample.

The second chapter investigates the impact of Ulaanbaatar’s 2019 raw coal ban on air pollution and child health using city- and province-level data from the National Statistics Office of Mongolia. Results from an event study framework indicate the ban reduced particulate air pollution (PM10) by 20%, but had the unintended consequence of increasing sulfur dioxide air pollution (SO2) by 87%. Effects of the ban on SO2 air pollution are robust to various clustering techniques and are more pronounced during winter months. While a policy goal was to improve health outcomes for infants and children under five, there is little evidence the ban altered mortality rates among these groups. Correlational evidence suggests the increase in ambient SO2 pollution may have offset potential mortality reductions from the decrease in ambient PM10 pollution.

The third chapter exploits the plausibly-exogenous timing of public transit strikes in London to demonstrate that strikes have health costs in the context of cyclist safety. Using a novel data set that records the geo-location of all known cyclist-involved crashes in London between 2014-2016, I show the first day of a transit strike causes a 31.9% increase in the expected number of cyclist-involved crashes in wards where the city has implemented bike-sharing stations. This increase in crashes likely operates through the mechanism of an unexpected increase in the number of cyclists during strikes, as transit strikes in London cause temporary substitution to the city’s bike-sharing system. The effect of strikes on cyclist-involved crashes is robust to multiple empirical specifications and comparison groups. Focusing on days when strikes occurred reveals heterogeneous effects of strikes on the number of crashes: strikes that impact the entire Tube network are associated with more crashes than strikes that only impact select transit lines