Replication Files: Crimes on the Move: The Effect of Ridesharing Services on Crime
Description
The rapid rise of ridesharing services like Uber and Lyft has transformed urban transportation, but their influence on crime and the mechanisms driving these changes remain under explored. This study examines the impact of the introduction of ridesharing services like Uber and Lyft on crime rates across U.S. cities, leveraging a natural experiment created by their staggered rollout. Using a Two-way Fixed Effects (TWFE) model, I find significant reductions in violent crimes, property crimes, and burglary following the entry of these services, while finding no substantial effects on larceny, motor vehicle theft, or arson. I also investigate the heterogeneity of these effects across different demographic groups, indicating that ridesharing services may influence crime through mechanisms such as employment and demographic changes. This research provides new evidence on the relationship between ridesharing services and crime, providing new insights into the potential mechanisms behind these effects.
Files
Steps to reproduce
This replication folder contains all the codes and cleaned data used to generate all figures and tables using Stata. The data cleaning process and construction of the final dataset are described in the data construction file. The raw data would be available upon request. This folder has the following three subfolders: I. Code: All the Stata codes in .do file format II. Data: All the cleaned data in .dta file format III. Outcome: Graphs, tables and latex files The following Stata scripts are included in the Code subfolder: • main_analysis.do: This Stata script produces the tables and figures for the main results in the paper. • mechanisms.do: This Stata script produces the tables for the “Mechanisms” section in the paper. • robustness.do: This Stata script produces the tables for the robustness analysis in the paper. • summary.do: This Stata script produces all the summary statistics tables in the paper. • Heterogeneity.do: This Stata script produces the tables for the “Heterogeneity” section in the paper. • graphical analysis.do: This Stata script produces all other figures in the paper. The following data are included in Data subfolder: • launch_city_code.dta: City level launch dates of Uber and Lyft since 2010 • ucr_uber.dta: Final cleaned dataset of UCR crime occurrences at the city level with the launch date of Uber and Lyft in each city. • ucr_uber_acs.dta: Final cleaned dataset of UCR crime occurrences at the city level with the American Community Survey controls and the launch date of Uber and Lyft in each city. • labor_uber.dta: Bureau of Labor Statistics Data at the city level with the launch date of Uber and Lyft in each city. • acs_msa_cleaned.dta: American Community Survey at the MSA level with the launch date of Uber and Lyft in each MSA (see data construction file for more details). • transportation_acs.dta: American Community Survey data on the use of mode of transportation. • ucr_uber_google.dta: Data of Google search intensity for the keyword “Uber” per U.S. city (see data construction file for more details). • nibrs_uber.dta: Final cleaned dataset of NIBRS crime occurrences at the city level with the launch date of Uber and Lyft in each city. All tables, figures, and LaTeX files needed to reproduce the paper are provided in the 'outcome' subfolder. Software: The code was last executed using Stata 4-core MP version 17.